Adaptive Single-Phase Reclosing in Transmission Lines Adaptive Single-Phase Reclosing in Transmission Lines

View metadata, citation and similar papers at core.ac.
uk brought to you by CORE

provided by Scholarship@Western
Western University
Scholarship@Western
Electronic Thesis and Dissertation Repository
11-11-2015 12:00 AM
Adaptive Single-Phase Reclosing in Transmission Lines

Farzad Zhalefar
The University of Western Ontario
Supervisor
Prof. Tarlochan Singh Sidhu
Graduate Program in Electrical and Computer Engineering

A thesis submitted in partial fulfillment of the requirements for the degree in Doctor of
Philosophy
© Farzad Zhalefar 2015
Follow this and additional works at: https://ir.lib.uwo.ca/etd
Part of the Power and Energy Commons, and the Signal Processing Commons
Recommended Citation
Zhalefar, Farzad, "Adaptive Single-Phase Reclosing in Transmission Lines" (2015). Electronic Thesis and
Dissertation Repository. 3377.
https://ir.lib.uwo.ca/etd/3377
This Dissertation/Thesis is brought to you for free and open access by Scholarship@Western. It has been accepted
for inclusion in Electronic Thesis and Dissertation Repository by an authorized administrator of
Scholarship@Western. For more information, please contact wlswadmin@uwo.ca.
ADAPTIVE SINGLE-PHASE RECLOSING IN TRANSMISSION LINES
(Thesis format: Monograph)
by
Farzad Zhalefar
Graduate Program in Electrical and Computer Engineering
A thesis submitted in partial fulfillment
of the requirements for the degree of
Doctoral of Philosophy
The School of Graduate and Postdoctoral Studies
London, Ontario, Canada
c Farzad Zhalefar 2015

Abstract
This research work is mainly concerned about dealing with temporary short circuit faults
in power system transmission lines. In fact, there are two types of electrical faults in power
systems, namely temporary and permanent. When a fault is permanent, the only way to clear
it is to de-energize the transmission line by opening the associated circuit breakers. However,
in many cases the fault is not solid and is caused by objects such as flying birds or broken
branches of trees. For these cases, electrical arc plays a major role. For such fault cases, it
is also possible to de-energize the faulted phase, temporarily, and re-energize it after a short
delay by reclosing the opened circuit breakers. This operation is called single-phase reclosing.
There is a chance that the fault becomes clear by natural extinction of the arc after the faulted
phase isolation in case the fault is temporary.
There are two considerable challenges regarding traditional single-phase reclosing in trans-
mission lines. The first challenge is the determination of the fault type, i.e., permanent or
temporary, as there is no guarantee that the fault is temporary. This is crucially important as
reclosing-on-fault, i.e., reclosing the opened breakers while the fault still stands, is harmful
for both power system stability and power system equipment. The second challenge which is
regarding temporary faults only, is that there is still no guarantee that the arc is extinguished
by the moment of reclosing. In such cases, reclosing leads in re-striking of arc and therefore,
an unsuccessful reclosing.
This research work is conducted in two phases. At the first phase, two adaptive methods
are developed to improve the traditional reclosing method upon the two challenges mentioned
in the second paragraph. The developed methods are capable of recognition of the fault type in
a reasonable amount of time after single-phase isolation of the line. Therefore, the protection
system will be able to block the reclosing function in case the fault is recognized as permanent
and to issue three-phase-trip signal as the next action. For temporary faults, re-energizing of the
ii
isolated phase by reclosing the opened breakers is the next action which has to be performed
after the arc extinction. The developed methods also have the capability of detection of the arc
extinction and therefore, a better performance for temporary fault cases is guaranteed. This is
the second feature required for an adaptive reclosing method.
The second phase of the research project is to estimate the arc extinction time well in
advance in case the fault is temporary. The idea is that three-phase tripping could be the right
action if the arc extinction time is too long as working under unbalanced conditions for an
unnecessarily long time duration is harmful for the power system.
Both of the proposed adaptive single-phase reclosing methods in this research work employ
local voltage information. Therefore, communication facilities are not needed for implemen-
tation of the proposed methods. It is shown in the thesis that the proposed methods are able to
quickly detect the fault type and also the arc extinction if the fault is temporary. Also, the two
proposed arc extinction time prediction methods are capable of prediction of the arc extinction
time well in advance and with acceptable precision.

All four proposed methods are effective for various system configurations including ideally-
transposed, untransposed and partially-transposed transmission lines and also for transmission
lines with different compensation conditions including with and without shunt reactor. Superior
performance of the proposed methods have been verified using 550 case studies simulated in
PSCAD and Matlab, and also a field recorded temporary fault case associated with a 765
kV transmission line. The 550 simulated case studies include 100 ideally-transposed, 240
untransposed and 210 partially-transposed line cases. The performances of the two proposed
reclosing methods are also compared with two of the existing adaptive reclosing methods where
considerable improvements are observed.
Keywords: Adaptive protection, arc extinction, single-phase reclosing, tranmission line
iii
To my parents ...
iv
Acknowledgments
I would like to express my sincere gratitude to Dr. Tarlochan Singh Sidhu, my Ph.D. super-
visor at Western University, for his helps during my studies. Also, I would like to show my
thankfulness to Dr. Mohammad Dadashzadeh whom I had the opportunity of working with at
Western. And at last but not least, whatever I have achieved would not be possible without all
the helps and supports from my dearest family.
v
Contents
Abstract ii
Dedication iv
Acknowledgments v
List of Figures x
List of Tables xviii
1 Introduction 1
1.1 Single-Phase Reclosing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Literature Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Objectives, Contributions and Novelties . . . . . . . . . . . . . . . . . . . . . 9
2 Reclosing in Transmission Lines 11

2.1 History of Reclosing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Single-Phase Reclosing in Canada . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3 Various Reclosing Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.1 High-speed vs. time-delayed . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Single-shot vs. multiple-shot . . . . . . . . . . . . . . . . . . . . . . . 17

2.3.3 Single-phase vs. three-phase . . . . . . . . . . . . . . . . . . . . . . . 18
vi
2.4 Protection System Operation During Reclosing . . . . . . . . . . . . . . . . . 20
2.5 Application Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 IEEE Guides for Auto-Reclosing . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.6.1 Fundamentals and applications . . . . . . . . . . . . . . . . . . . . . . 29
2.6.2 Auto-reclosing for transmission systems . . . . . . . . . . . . . . . . . 32

2.6.3 Auto-reclosing methods . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.6.4 Auto-reclosing settings . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.6.5 Auto-reclosing blocking . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.6.6 Operation considerations . . . . . . . . . . . . . . . . . . . . . . . . . 38
3 Modeling and Behavior Analysis During Single-phase Reclosing 54
3.1 Modeling of the Studied System . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1.1 Power system modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1.2 Arc modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1.3 Protective relay modeling . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2 Behavior Analysis of the System During Single-Phase Reclosing . . . . . . . . 66
3.2.1 Transposed line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Absence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 66
Presence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.2.2 Untransposed line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Absence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 72
Presence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4 Fault Type Recognition and Arc Extinction Detection 76
4.1 Fundamentals of the Reclosing Methods . . . . . . . . . . . . . . . . . . . . . 77
4.1.1 Angle-based analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
vii
Ideally-transposed line . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Untransposed line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Partially-transposed line . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.1.2 Phasor-based analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Permanent fault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Temporary fault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Arc extinction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.2 Proposed Adaptive Reclosing Algorithms . . . . . . . . . . . . . . . . . . . . 100
4.2.1 Angle-based algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Fault type recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Arc extinction detection . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.2 Phasor-based algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Fault type recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Arc extinction detection . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.3 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3.1 Angle-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3.2 Phasor-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4.3.3 Comparison of the reclosing methods . . . . . . . . . . . . . . . . . . 124
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5 Arc Extinction Time Prediction 127
5.1 Fundamentals of The Proposed Prediction Methods . . . . . . . . . . . . . . . 128
5.1.1 Analysis of the faulted phase voltage waveforms . . . . . . . . . . . . 128
5.2 Proposed Prediction Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.2.1 Point-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Exponential curve approximation for the point-based method . . . . . . 133
viii
Proposed algorithm for the point-based method . . . . . . . . . . . . . 133
5.2.2 Window-based method . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Exponential curve approximation for the window-based method . . . . 135
Proposed algorithm for the window-based method . . . . . . . . . . . . 136
5.3 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

5.3.1 Point-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.3.2 Window-based method . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.3.3 Comparison of the prediction methods . . . . . . . . . . . . . . . . . . 153
5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6 Conclusions 155
Bibliography 159
ix
List of Figures
2.1 Sequence of operation of the protective relay during a successful reclosing [74]. 21
2.2 Sequence of operation of the protective relay during an unsuccessful reclosing

[74]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1 Single-line diagram of the modeled power system. . . . . . . . . . . . . . . . . 57
3.2 Single-line diagram of a power system with ideally-transposed and untrans-

posed transmission lines simulated in PSCAD. . . . . . . . . . . . . . . . . . . 59
3.3 Single-line diagram of a power system with partially-transposed transmission
line simulated in PSCAD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.4 Details of the arc model simulated in PSCAD. . . . . . . . . . . . . . . . . . . 65
3.5 (a) Equivalent circuit of a transposed line without shunt reactor in case of
single-phase reclosing of phase s, (b) rearranged circuit. . . . . . . . . . . . . . 67
3.6 Normalized (a) magnitude and (b) phase angle of the faulted phase voltage. . . 69
3.7 (a) Equivalent circuit of a transposed transmission line with four-legged shunt
reactor, (b) simplified circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.8 (a) Equivalent circuit of an untransposed transmission line without shunt reac-
tor, (b) simplified circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.9 (a) Equivalent circuit of an untransposed transmission line with four-legged

shunt reactor, (b) rearranged circuit, (c) simplified circuit. . . . . . . . . . . . . 74
x
4.1 Electric arc voltage and current for the case of a temporary fault in an ideally-
transposed transmission line, in absence of shunt reactor; (b) is zoomed in (c). . 77
4.2 Electric arc voltage and current for the case of a temporary fault in an ideally-
transposed transmission line, in presence of shunt reactor; (b) is zoomed in
(c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.3 Electric arc voltage and current for the case of a temporary fault in an untrans-
posed transmission line, in absence of shunt reactor; (b) is zoomed in (c). . . . . 78
4.4 Electric arc voltage and current for the case of a temporary fault in an untrans-
posed transmission line, in presence of shunt reactor; (b) is zoomed in (c). . . . 79
4.5 Electric arc voltage and current for the case of a temporary fault in a partially-
transposed transmission line, in absence of shunt reactor; (b) is zoomed in (c). . 79
4.6 Electric arc voltage and current for the case of a temporary fault in a partially-
transposed transmission line, in presence of shunt reactor; (b) is zoomed in
(c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.7 (a) δmean for the case of a permanent fault in an ideally-transposed transmission
line, in absence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . 82
4.8 (a) δmean for the case of a permanent fault in an ideally-transposed transmission
line, in presence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . 83
4.9 (a) δmean for the case of a temporary fault in an ideally-transposed transmission
4.10 (a) δmean for the case of a temporary fault in an ideally-transposed transmission
4.11 (a) δmean for the case of a permanent fault in an untransposed transmission line,
in absence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . . . . 85
4.12 (a) δmean for the case of a permanent fault in an untransposed transmission line,
in presence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . . . 86
xi
4.13 (a) δmean for the case of a temporary fault in an untransposed transmission line,
in absence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . . . . 86
4.14 (a) δmean for the case of a temporary fault in an untransposed transmission line,
in presence of shunt reactor; (a) is zoomed in (b) and (c). . . . . . . . . . . . . 87
4.15 (a) δmean for the case of a permanent fault in a partially-transposed transmission
4.16 (a) δmean for the case of a permanent fault in a partially-transposed transmission
4.17 (a) δmean for the case of a temporary fault in a partially-transposed transmission
4.18 (a) δmean for the case of a temporary fault in a partially-transposed transmission
4.19 Normalized faulted phase voltage magnitude and its first derivative for a perma-
nent fault case in an ideally-transposed transmission line, in absence of shunt
reactor; (b) is zoomed in (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.20 Normalized faulted phase voltage magnitude and its first derivative for a per-
manent fault case in an untransposed transmission line, in presence of shunt
4.21 Normalized faulted phase voltage magnitude and its first derivative for a perma-
nent fault case in a partially-transposed transmission line, in absence of shunt

4.22 Normalized faulted phase voltage magnitude and its first derivative for a tem-
porary fault case in an ideally-transposed transmission line, in absence of shunt
xii
4.23 Normalized faulted phase voltage magnitude and its first derivative for a tempo-
rary fault case in an ideally-transposed transmission line, in presence of shunt
rary fault case in an untransposed transmission line, in absence of shunt reactor;
(b) is zoomed in (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
porary fault case in an untransposed transmission line, in presence of shunt
porary fault case in a partially-transposed transmission line, in absence of shunt
rary fault case in a partially-transposed transmission line, in presence of shunt
4.28 First angle derivative for the case of a temporary fault case in an untransposed
transmission line, in absence of shunt reactor; (a) is zoomed in (b). . . . . . . . 99
4.29 First angle derivative for the case of a temporary fault case in an untransposed
transmission line, in presence of shunt reactor; (a) is zoomed in (b). . . . . . . . 100
4.30 Block diagram of the proposed angle-based adaptive single-phase reclosing

algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.31 Block diagram of the proposed adaptive single-phase reclosing algorithm. . . . 104
4.32 Average temporary fault detection time for an ideally-transposed and an un-
transposed transmission line in Absence (A) and Presence (P) of shunt reactor
using the angle-based method (ms). . . . . . . . . . . . . . . . . . . . . . . . . 110
xiii
4.33 Average permanent fault detection time for an ideally-transposed and an un-
transposed transmission line in Absence (A) and Presence (P) of shunt reactor
4.34 Average temporary fault detection time for a partially-transposed transmission

line in Absence (A) and Presence (P) of shunt reactor using the angle-based
method (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.35 Average permanent fault detection time for a partially-transposed transmission
line in Absence (A) and Presence (P) of shunt reactor using the angle-based
method (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.36 Average reclosing time delay for an Ideally-transposed (I) and an Untransposed
(U) transmission lines in Absence (A) and Presence (P) of shunt reactor using
the angle-based method (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.37 Maximum reclosing time delay for an Ideally-transposed (I) and an Untrans-
posed (U) transmission lines in Absence (A) and Presence (P) of shunt reactor
4.38 Average reclosing time delay for a partially-transposed transmission line in

Absence (A) and Presence (P) of shunt reactor using the angle-based method
(ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.39 Maximum reclosing time delay for a partially-transposed transmission line in

Absence (A) and Presence (P) of shunt reactor using the angle-based method
(ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.40 Waveforms of a temporary fault case in Rockport 765 kV station in American

Electric Power grid, (a) normalized voltage magnitude of the faulted phase and
(b) average value of the faulted phase voltage angle; (b) is zoomed in (c) and (d).118
xiv
4.41 Average reclosing time delay for an Ideally-transposed (I) and an Untransposed
(U) transmission lines in Absence (A) and Presence (P) of shunt reactor using
the phasor-based method (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.42 Maximum reclosing time delay for an Ideally-transposed (I) and an Untrans-
posed (U) transmission lines in Absence (A) and Presence (P) of shunt reactor
using the phasor-based method (ms). . . . . . . . . . . . . . . . . . . . . . . . 121
4.43 Average reclosing time delay for a partially-transposed transmission line in

Absence (A) and Presence (P) of shunt reactor using the phasor-based method
(ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.44 Maximum reclosing time delay for a partially-transposed transmission line in

Absence (A) and Presence (P) of shunt reactor using the phasor-based method
(ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.45 Waveforms of a temporary fault case in Rockport 765 kV station in Ameri-

can Electric Power grid, (a) normalized instantaneous voltage waveform of the
faulted phase, (b) normalized voltage magnitude of the faulted phase and (b)
its first derivative, (c) first angle derivative. . . . . . . . . . . . . . . . . . . . . 123
5.1 Electric arc voltage and current for the case of two temporary faults in an
ideally-transposed transmission line in {(a) and (c)} absence and {(b) and (d)}
presence of shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively. . . 129
5.2 Electric arc voltage and current for the case of two temporary faults in an un-
transposed transmission line in {(a) and (c)} absence and {(b) and (d)} presence
of shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively. . . . . . . . 129
5.3 Electric arc voltage and current for the case of two temporary faults in a partially-
transposed transmission line in {(a) and (c)} absence and {(b) and (d)} presence
of shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively. . . . . . . . 130
xv
5.4 Normalized faulted phase voltage magnitude and its first derivative for the tem-
porary fault cases in an ideally-transposed transmission line shown in Figure
5.1; (c) and (d) are zoomed in (e) and (f), respectively. . . . . . . . . . . . . . . 131
porary fault cases in an untransposed transmission line shown in Figure 5.2; (c)
and (d) are zoomed in (e) and (f), respectively. . . . . . . . . . . . . . . . . . . 131
porary fault cases in a partially-transposed transmission line shown in Figure
5.3; (c) and (d) are zoomed in (e) and (f), respectively. . . . . . . . . . . . . . . 132
5.7 V snD and the approximating exponentials associated with the temporary fault
cases shown in Figure 5.5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.8 V snD curves shown in (a) Figure 5.4-(e), (b) Figure 5.4-(f), (c) Figure 5.5-(e),
(d) Figure 5.5-(f), (e) Figure 5.6-(e) and (f) Figure 5.6-(f), and the associated
approximating exponentials. . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.9 Average arc extinction prediction time error of the point-based method for the
ideally-transposed transmission lines in Absence (A) and Presence (P) of shunt
reactor (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
ideally-transposed transmission lines in Absence (A) and Presence (P) of shunt
reactor (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
untransposed transmission lines in Absence (A) and Presence (P) of shunt re-
actor (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
untransposed transmission lines in Absence (A) and Presence (P) of shunt re-
actor (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
xvi
5.13 Average arc extinction prediction time error of the point-based method for a
partially-transposed transmission line in absence of shunt reactor (ms). . . . . . 144
partially-transposed transmission line in absence of shunt reactor, in percent. . . 144
partially-transposed transmission line in presence of shunt reactor (ms). . . . . 145
partially-transposed transmission line in presence of shunt reactor, in percent. . 145
5.17 (a) Normalized instantaneous waveform, (b) magnitude and (c) the first deriva-
tion of the faulted phase voltage, (d) the approximating exponential curve of
the 765 kV faulted phase using the point-based method. . . . . . . . . . . . . . 146
5.18 Average arc extinction prediction time error of the window-based method for an
Ideally-transposed (I) and an Untransposed (U) transmission lines in Absence
(A) and Persence (P) of shunt reactor (ms). . . . . . . . . . . . . . . . . . . . . 150
5.19 Average arc extinction prediction time error of the window-based method for an
Ideally-transposed (I) and an Untransposed (U) transmission lines in Absence
(A) and Persence (P) of shunt reactor (%). . . . . . . . . . . . . . . . . . . . . 150
5.20 Average arc extinction prediction time error of the window-based method for a
partially-transposed transmission line in Absence (A) and Presence (P) of shunt
reactor (ms). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.21 Average arc extinction prediction time error of the window-based method for a
partially-transposed transmission line in Absence (A) and Presence (P) of shunt
reactor (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.22 Normalized instantaneous waveform, magnitude, the first derivation of the faulted
phase voltage and the approximating exponential curve for the real temporary
fault case using the window-based method. . . . . . . . . . . . . . . . . . . . . 152
xvii
List of Tables
3.1 Modeled Arc Parametrs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.1 Average temporary fault detection time for an ideally-transposed and an un-
transposed line using the angle-based method (ms). . . . . . . . . . . . . . . . 108
4.2 Average permanent fault detection time for an ideally-transposed and an un-
transposed lines using the angle-based method (ms). . . . . . . . . . . . . . . . 108
4.3 Average fault type recognition time for a partially transposed transmission line
4.4 Average reclosing time delay for an ideally-transposed and an untransposed

lines using the angle-based method (ms). . . . . . . . . . . . . . . . . . . . . . 113
4.5 Maximum reclosing time delay for an ideally-transposed and an untransposed

lines using the angle-based method (ms). . . . . . . . . . . . . . . . . . . . . . 114
4.6 Average and maximum reclosing time delay for a partially transposed trans-
mission line using the angle-based method (ms). . . . . . . . . . . . . . . . . . 114
4.7 Average/maximum reclosing time delay for an ideally-transposed and an un-

transposed lines using the phasor-based method (ms); Sh.R.: Shunt Reactor;
S.C./M.C.: Fault on Side/Middle Conductor. . . . . . . . . . . . . . . . . . . . 119
4.8 Average and maximum reclosing time delay for a partially transposed trans-
mission line using the phasor-based method (ms); Sh.R.: Shunt Reactor. . . . . 120
xviii
5.1 Average arc extinction prediction time error of the point-based method for an
ideally-transposed/untransposed transmission line (ms). . . . . . . . . . . . . . 139
5.2 Average arc extinction prediction time error of the point-based method for an
ideally-transposed/untransposed transmission line in percent of the real arc ex-
tinction time (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

partially-transposed transmission line (ms). . . . . . . . . . . . . . . . . . . . 140
partially-transposed transmission line in percent of the real arc extinction time
(%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.5 Average arc extinction prediction time error of the window-based method for
an ideally-transposed transmission line in (ms) and percent of the real arc ex-
tinction time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
an untransposed transmission line in (ms) and percent of the real arc extinction
time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
a partially-transposed transmission line in (ms) and percent of the real arc ex-
tinction time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
xix
Chapter 1
Introduction
Power system is one of the most important systems developed by humankind. Basic role of this
system is to produce the electrical energy in power plants, etc., transfer it to the desired desti-
nations and finally, distribute it among consumers. Each of these steps has its own importances
and difficulties. In places in which locations of electrical energy production and consumption
are far from each other, transferring the energy by means of transmission lines becomes ex-
tremely critical. One of the factors that can disturb the ideal performance of the transmission
lines is short circuit fault inception against which the lines must be protected [1].
This research work is dedicated to adaptive single-phase reclosing in transmission lines. In

the performed research, adaptive methods are proposed for fault type recognition purpose, i.e.,
permanent and/or temporary, at first, and then making a proper decision based on the detected
fault type. The proper decision is always to initiate three-phase trip signal for permanent fault
cases. For cases of temporary fault, normally, the correct decision is to reclose the breakers
after the arc extinction detection. However, if the arc is being extinguished too slowly, it can
be a better action to issue the three-phase trip signal and avoid reclosing as it is harmful for the
system to be kept in unbalanced conditions for a long time. For this purpose, arc extinction
time must be predicted in advance. This is also covered in the performed research work.
1
2 Chapter 1. Introduction
In this Chapter, at first, there is an introduction about single-phase reclosing followed by

problem definition and literature survey sections. Finally, objectives, contributions and novel-
ties of the performed research study are presented.
1.1 Single-Phase Reclosing
According to statistics, most of the faults happening in extra and ultra high voltage transmission
lines are of temporary nature [1, 2]. When dealing with temporary faults, which are mostly
accompanied by electric arcs, one interesting idea is to isolate the faulted line, temporarily, and
re-energize it with or without a delay. This operation is called reclosing [1, 3].
The fastest way to extinguish an arc associated with a temporary fault is three-phase re-
closing. This is because of the fact that after isolation of all three conductors, there would
be no voltage source to feed the arc. However, this could be harmful for the stability of the
system as two sides of the line will remain disconnected for a considerable amount of time. If
the temporary fault type is single-phase-to-ground, a good trade off is single-phase reclosing
[3, 4]. This technique can be very practical as faults are mostly single-phase-to-ground [1].
In single-phase reclosing, the two healthy phases are kept in circuit during the fault period
while the faulted phase becomes de-energized, temporarily, by opening operation of the circuit
breakers. In this technique, the faulted phase will be returned to service ideally when the fault
is cleared [3]. Compared to three-phase reclosing, the problem with single-phase reclosing
technique is that the arc can not be extinguished, immediately, as it will be fed through mutual
coupling with the non-faulted phases [2].
The critical point about reclosing is to avoid reclosing-onto-fault, i.e., reclosing when the
fault is not cleared yet. This can destroy the equipment and put the system stability at risk [1].
Therefore, the reclosing system has to wait for the arc to be extinguished in case the fault is
temporary.
1.2. Problem Definition 3
1.2 Problem Definition
In a traditional single-phase reclosing method, single-phase isolation of the faulted phase is

performed as soon as the single-phase-to-ground fault is detected. The next step is to issue
single-phase reclosing signal after a predetermined delay called dead-time [2]. There are two
main problems with the traditional reclosing method. First, there is no guarantee that the fault
type is temporary. In fact, traditional method recloses the faulted phase regardless of the fault
type. Second, even if the fault type is temporary, there is no guarantee that the reclosing be
performed after the arc extinction. This is a very big issue because reclosing prior to the arc
extinction will lead to restriking of the arc which means the reclosing is not successful.
The next problem is regarding only temporary fault cases. For permanent fault cases, the
proper decision is definitely to initiate the three-phase trip signal. But for temporary fault
cases, the reclosing method must be able to detect the arc extinction time and the reclosing
signal must be initiated after the arc extinction detection. However, arc extinction time can be
too long for some temporary fault cases. Therefore, the proper decision can be to initiate the
three-phase trip signal for such cases.
Also, regarding the proposed methods, the preference is that the methods utilize local volt-
age measurements information to perform reclosing as communication facilities are not avail-
able all the times. In addition, the proposed methods must be valid for transmission lines with
any transposition and compensation conditions.
1.3 Literature Survey
According to statistics, more than 80% of faults in overhead transmission lines are tempo-
rary, mostly single-phase-to-ground faults [2]. There are different reasons for temporary fault
inceptions including overvoltages as results of lightnings and temporary contacts between
phase/ground wires, e.g., by broken branches of trees or flying birds. As these faults are not per-
manent and can be cleared by themselves, one practical solution is single-phase auto-reclosing,
SPAR, of the faulted phase [5]. This way, there is a chance to bring the faulted phase back to
operation shortly after the fault inception which leads to enhancement of reliability and stabil-
ity of the system [6]. Even during single-phase opening of the faulted phase breaker, still 58%
of the line capacity available due to presence of two non-faulted phases [7].
Traditionally, when SPAR function is considered, reclosing is performed after single-phase

opening of the faulted phase by a predetermined time delay called dead-time [3, 8]. In such a
condition, first, reclosing is performed regardless of the fault type, i.e., permanent or temporary.
Second, there is no guaranty if the arc is extinguished at the moment of reclosing for temporary
fault cases. Therefore, there is a danger of reclosing-onto-fault for permanent fault cases and
restriking of the arc for temporary faults in which the arc is not extinguished at the time of
reclosing. Both these scenarios which are considered as unsuccessful reclosing attempts,are
dangerous for the power system and the system equipment [9]-[12]. Additionally, reclosing
process is usually repeated after a short break in traditional SPAR if unsuccessful which makes
the situation even worse [2], [13]-[15]. In [16]-[19] destructive effects of single-phase reclosing
on torsional torques of turbine-generator shaft are analyzed and discussed.
There have been some methods proposed and implemented in practice to minimize the
dead-time including installation of three-phase four-legged shunt reactor with inductively groun-
ded neutral at both ends of transmission lines [2, 20, 21]. By properly selection of the neutral
reactor, it is possible to limit the voltage across the arc and therefore, reduce the dead-time.
This method is more effective for transposed transmission lines, although it has been applied to
untransposed lines at very high voltage levels using switching scheme of shunt reactor at one
end [2, 21, 22].
To overcome the mentioned problems regarding the traditional SPAR, several adaptive
auto-reclosing methods are proposed. The main objectives of an adaptive single-phase re-
1.3. Literature Survey 5
closing technique is to quickly identify permanent fault and to detect the arc extinction in cases
of temporary fault [2, 23, 24]. Most of the proposed adaptive SPAR methods in the litera-
ture use some features of the voltage/current waveforms [25]-[30]. In [25] 27 real temporary
fault cases in 550 kV transmission lines are analyzed based on the voltage wave shapes. Zero-
crossing point of the voltage is employed for enhancing the reclosing performance in [26]. In
[27] and [28] frequency characteristics of the current are used for reducing the reclosing dead
time. In [31], a method for arc extinction detection is proposed based on the behavior of the
faulted phase voltage. The proposed method is also verified using test results from BC Hydro
network. This paper in which information of only one side of the transmission line is employed
for arc extinction detection, is a sample of contribution of Canadian researchers in the area of
single-phase reclosing.
In the method presented in [32] which is registered as a US patent, the exact moment of
arc extinction is detected by measuring the characteristics of the faulted phase voltage. In this
method, depending on the compensation level of the line, one of the prepared algorithms is
chosen for this purpose. Using this method, reclosing-on-fault will be prevented. In another
patent, three-phase reclosing of shunt reactor is considered [33]. Three-phase voltages and
currents are used for detection of the arc extinction time in the patent registered as [34].
Many researchers have investigated techniques for enhancing the effect of reclosing on
system stability [36]. Recently, there have been major research activities to deal with this
problem. As an example, artificial neural networks, ANNs, have been employed to detect the
secondary arc extinction time after the breaker opening [37]-[39]. In [37] and [38], ANN is
employed to extract appropriate features of frequency components of the voltage. Neuro-Prony
and Taguchis methodology are used in [39] for a better reclosing performance. Also, wavelet
analysis and neural network based SPAR schemes have been developed to recognize certain
situations in order to decide if to reclose the circuit breakers [40]. However, the application of
these techniques requires a broad data from the faulted system for the neural network training.
Fuzzy logic is employed in [30] and [41] for increasing the stability margin of the power
system by enhancing the reclosing performance of the circuit breakers. In [42] a Fuzzy logic-
based method is proposed for enhancing the single-phase reclosing in double-circuit transmis-
sion lines. This method identifies the fault type based on the angular difference between the
positive and negative sequence components of the voltage. Reference [43] is specifically fo-
cused on secondary arc extinction in single-phase reclosing. In this paper, the role of neutral
reactor as a tool for having a better arc extinction performance is of a high importance.
In paper [44] variable dead-time SPAR scheme has been presented based on stability mar-
gin. This scheme which is applicable only when the power system stability degree is suf-
ficiently high, increases the reclosing dead-time [6]. In [45], voltage pattern is employed for
reduction of the reclosing dead-time. Paper [46] proposes a method for decrement of the reclos-
ing dead-time by reducing the mutual capacitive coupling effect. For this purpose, a capacitor
in parallel to each pole of the circuit breakers must be considered.
One of the ways of categorizing adaptive SPAR methods is based on decision making
method. In fact, all methods extract some features from waveforms or spectrum of the faulted
system variables, including power, current or voltage [47]-[48]. These features are either com-
pared to some thresholds or used independently for decision making.
There are many examples of methods use threshold for decision making. In [49, 50] zero
sequence power as well as the faulted phase voltage are employed for arc extinction detection.
In [52]-[54], spectrum of the faulted phase current is used for identification of the fault type.
The method proposed in reference [6] uses rms value of the faulted phase voltage for the fault
type recognition. For this purpose, rms value is calculated for different time periods and when
the difference between rms values is larger than a predetermined threshold, arc extinction is
detected. Arc extinction detection is performed in [54] using total harmonic distortion, THD,
of the faulted phase voltage. This method works based on the fact that the amount of harmonic
distortion of the voltage considerably decreases after the secondary arc extinction.
1.3. Literature Survey 7
There are also few methods proposed based on the harmonic characteristics of the faulted
phase voltage after single-phase opening of the breakers [55]. These methods are applicable
only to uncompensated transmission lines. The reason is that the shunt reactor resonates with
the line capacitances which leads to appearance of sub-synchronous oscillatory voltage wave-
forms in the faulted phase while the breaker is open. This can lead to inaccuracy in harmonic
estimation and therefore, error in arc extinction detection [2]. There are more examples in [6],
[56]-[61] in this order. However, the main problem with most of the methods of both groups
is that these methods mostly depend on the under study system and the associated settings will
change based on the parameters of the system [55].
Fourier transform, either fast Fourier transform, FFT, or discrete Fourier transform, DFT,
is the most practical technique for analysis of the current and voltage waveforms [2], [58].
Other options can be few techniques such as wavelet [40], [59]. Although by using different
mother wavelets, more data can be extracted, sampling rate and the level of complexity will
also increase which is not appropriate and practical for protection and relaying. Therefore,
Fourier based methods are more accepted, practically [5].
Another way of categorizing adaptive SPAR methods is the data type usage. In other words,
generally there are two possible categories of reclosing methods available, namely local and
communication-based. Local methods use only the information of one side of the transmission
line while communication-based methods have access to data of both sides of the line. The
method proposed in [56] is an example of communication based methods which predicts the
faulted phase voltage after the arc extinction using the synchronized voltage phasor of the other
healthy phases from both ends of the transmission line. Therefore, the instant of arc extinction
can be detected when the faulted phase voltage magnitude and angle are close enough to their
predicted values. However, this method is developed for arc extinction detection in ideally-
transposed lines and is not applicable to untransposed or partially transposed lines.
Communication-based reclosing methods are faster and more accurate that local methods,
but, they are not beneficial when communication facilities are not considered in that part of
the power system. These methods are also dependent on the communication facilities which
affects the reliability of the protection system. Therefore, local methods are more accepted in
practice [5]. Reference [62] is an example of adaptive SPAR methods which is based on local
voltage measurements. In this method, the faulted phase voltage magnitude is used for finding
a reference time to start the required calculations. Then, the combination of the magnitude and
phase angle is employed for detection of the fault type. Finally, arc extinction is detected based
on behavior of the phase angle of the induced voltage for temporary fault cases.
Controlled switching of breakers for the purpose of reduction of surge overvoltages in shunt
compensated transmission lines is proposed in [63]. In this method, unsuccessful reclosing is
also prevented as the proposed method is designed to detect the arc extinction time as well.
This method is also effective for both single-phase and double-phase-to-ground faults.
In [64], issues of single-phase and three-phase tripping of shunt reactor during reclosing are
discussed. The main source of the problem is the resonance between the shunt reactor and the
line capacitors. For solving the problem, it is proposed to switch off the shunt reactors before
the arc extinction. Therefore, there will be no resonance after the arc extinction. However, this
method can lead to some other issues and difficulties. The method proposed in the paper is to
use power resistors in the neutral of shunt reactors for damping the resonance. However, what
is missed in this analysis is the effect of neutral reactor which is usually used for a better arc
extinction performance.
During single-phase reclosing of lines, the voltage condition is unbalanced as one of the
conductors is isolated and only two phases are energized. In [65] application of Zig-Zag trans-
former for maintaining the balanced load voltages during single-phase reclosing is explained.
In this method, for the purpose of reducing the negative-sequence and zero-sequence voltages
of the load, application of a capacitor in the neutral of the Zig-Zag transformer is proposed.
1.4. Objectives, Contributions and Novelties 9
At the end, there is a different idea for adaptive single-phase reclosing called hybrid re-
closing scheme proposed in [7, 82]. In this method, the faulted phase is tripped, initially,
thereafter, the remaining two phases are tripped with a short delay. After the three-phase open-
ing, the residual energy in charged line capacitances and inductances feeds the arc until the
energy is consumed and the arc is quenched. In this case, if the fault is temporary and arc is
extinguished, a sinusoidal signal with a nonzero dc offset appears in the faulted phase volt-
age which can be considered as a sign of arc extinction, whereas if the fault is permanent,
the faulted phase voltage becomes zero and a permanent fault can be differentiated from the
temporary fault [2].
1.4 Objectives, Contributions and Novelties
In a traditional single-phase reclosing, the reclosing task is activated with a predetermined time
delay after the single-phase isolation of the line called dead-time. There are two main issues
about the traditional reclosing method. The first issue is that reclosing is performed regardless
of the fault type, i.e., permanent or temporary. The second issue is that even if the fault is
temporary, there is no guarantee that the secondary arc is extinguished before the reclosing.
In this research work, both of the issues of the traditional reclosing method are covered. The
third part of the research work is to propose a method for prediction of the arc extinction time in
case the fault is temporary. This is important as in some temporary fault cases the arc extinction
time is too long. Therefore, it can be reasonable to issue the three-phase trip signal instead of
reclosing despite the fault type.
The first part of the research study is dedicated to adaptive single-phase reclosing. For this
purpose, two reclosing methods are proposed. The first contribution of the proposed adaptive
single-phase reclosing methods is to quickly identify the fault type, i.e., permanent or tem-
porary. This is important as adaptive reclosing must be performed only if the fault type is
identified as temporary. In fact, reclosing the open breakers when the fault is permanent is
extremely dangerous for the power system equipment as well as the stability of the system.
The second contribution of the performed research study is to detect the arc extinction mo-
ment for temporary fault cases. In adaptive single-phase reclosing techniques, reclosing takes
place only when there are evidences for the arc extinction. Therefore, in cases of temporary
fault detection, the reclosing time delay is flexible depending on the situation [2].
There are some works done in the area of fault type recognition and arc extinction detec-
tion. The contribution of this research work is to develop faster and more reliable methods
compared to the existing techniques. Also, the developed methods are effective for any kind
of transmission line configurations, including ideally-transposed, untransposed and partially-

transposed, and with and without shunt reactors. The developed methods utilize local voltage
measurement data as communication facilities between the two sides of the transmission lines
are not always available.
Finally, regarding the arc extinction time prediction, there are no publications presented in
the literature in this area. In this research work, there are two adaptive methods proposed for
the purpose of prediction of arc extinction time. Therefore, the proposed prediction methods
are considered as novelty in single-phase reclosing research area.
Chapter 2
Reclosing in Transmission Lines
According to an IEEE Power Systems Relaying Committee report, auto-reclosing of circuit

breakers in transmission lines has been considered as an applicable and beneficial function in
power systems [5]. The main purpose of auto-reclosing is to keep the faulted transmission line
in service as long as possible which makes the power system more reliable and economical.
This concept works only when the fault is of temporary nature, i.e., faults in which the fault
resistance is not solid and it can be cleared by itself. Faults occurred due to a broken branch of
a tree is an example of temporary faults.
In reality, the idea of reclosing has been a very beneficial practice as according to statistics,
80% of faults in overhead transmission lines are of temporary nature [2]. The performance of
auto-reclosing task can be confirmed during disturbances including abnormal weather condi-
tions such as storms and heavy rains. In such conditions, the temporarily faulted system can
go back to operation without any special interruption and any need for maintenance [2].
In this Chapter, history of reclosing is discussed to see how this concept has been adopted
by power system utilities. Afterwards, application of single-phase reclosing in Canadian grids

is considered. Then, various reclosing methods in terms of reclosing speed and number of shots
and phases are described and the application of reclosing in various conditions are discussed.
11
12 Chapter 2. Reclosing in Transmission Lines
Later, the operation sequence of protection system during single-phase reclosing is explained
and at the end of this Chapter, IEEE guides for auto-reclosing including the settings and various
operating modes are explained. It must be noted that in order to stay consistent, protective
device in this Chapter means a device that performs the auto-reclosing task. This can be done
either simply as a logical circuit in the same device or as a separate device.
2.1 History of Reclosing
Beginning of the twentieths century was when reclosing used for the first time in radial feeder
circuits in which fuses and over-current relays had been employed for protection of the distri-
bution system. Studies showed that the initial version of reclosing was successful in 73 to 88
percent of the cases [2].
Inverse-time relays with instantaneous trip elements were introduced to the power sys-
tem in early 1930’s. These relays helped coordination with fuse schemes. At those days,
auto-reclosing techniques used for reclosing of the circuit following a predetermined delay for
deionization of the arc path and mechanical reset of the relay, only one time, and if relay trips
within 30 seconds after the first trip, lockout is considered. The Continuity of the service was
the only purpose of the first reclosing techniques.
Later, transmission level circuit breaker were introduced to the power system with high-
speed mechanical performances. Fault clearance time was reduced by faster operation of the
newly developed breakers. The faster operating speeds of these new circuit breakers reduced
clearing time, permitted high-speed reclosing by which system stability was also enhanced.
Also, minimum reclosing time was determined by studying of the flash-over probability in
insulators. This ensured the needed time for deionization of the arc path.
As the first, auto-reclosing applications for line with more than two terminals, in order to
circuit completion, it was practiced to perform a delayed reclosing at all line terminals provided
2.2. Single-Phase Reclosing in Canada 13
that the first reclosure be successful. Still the best reclosing practice from the continues serving
point of view is multi-shot reclosure.
In recent years, transformer failures caused by reclosing of distribution substation feeders
have concerned the system activists. Also, there is a chance of increasing stress on shafts
between generators and turbines. It must be noted that this issue has been discussed in a paper
at around 1940. It was concluded that despite transient power limits, single-pole switching
might be dictated.
Considerable enhancements in mechanism, designs, operation reliability and speed of high

voltage circuit breakers is been achieved, recently. Today, such achievements in addition to
developments in protection devices have made high-speed reclosing a practical scheme.
Due to nowadays achievements in solid state and digital technologies, one reclosing Relay
per Breaker has been considered again. e.g., one specified logic unit or reclosing processor can
be used for programming for each of the breakers, differently. This processor can check the
de-energized or energized status of all lines, buses, positions of all circuit breaker selector and
control switches, closing control voltages, circuit breaker positions, etc. As various stations
became available through communication, the condition knowledge at other substations can be
more local than central.
Having access to fast communication systems and data acquisition, now it is possible to
make control of the transmission line circuit breaker, central. But it seems that control of
reclosing systems must stay local at the substation level.
2.2 Single-Phase Reclosing in Canada
Generally speaking, single-phase reclosing is in practice mostly in Europe and Asia [25]. In
many North American provinces and states, single-phase reclosing is disabled due to the danger
of unsuccessful operation of traditional reclosing methods.
Single-phase reclosing has been in practice in BC Hydro transmission lines with four-
legged shunt reactors for the last 40 years [31]. Before 1970, high voltage transmission lines
in BC were not equipped with four-legged shunt reactors. In mid 1970s, BC Hydro started to
equip the new 500 kV lines with four-legged shunt reactors and applied single-phase reclosing
to the lines. Currently, in BC Hydro grid, single-phase reclosing is disabled for 500 kV lines
without shunt reactor [66].
According to Alberta interconnected electric system protection standard, all 240 kV and
higher transmission lines shall be equipped with single-phase reclosers [67]. The minimum
value for dead time setting is 0.75 s in Alberta. For longer dead time durations, appropriate
system studies must be performed while shorter dead-time durations are not allowed [66]. In
Quebec, 330 kV transmission lines are equipped with single-phase reclosers with dead time
settings of 0.75 s and 1 s while this is not applicable to 735 kV lines [68]. Single-phase
reclosing is applicable for 500 kV lines in Manitoba [69]. In Ontario, reclosing is applicable
for distribution networks [70].
A different reclosing technique called single-phase switching obtained by combining single-

phase reclosing and single-phase tripping was put in practice on 155 kV and 220 kV European
transmission lines in 1950s [71]. The main reason of application of single-phase switching was
to increase the power transfer capacity of the line and not the stability issues. This methods
was applicable for transmission lines shorter than 40 km in length with no ground wires. This
is because in such lines single-phase faults are very common while secondary arc extinction is
not an issue. This technology was extended to 420 kV lines in 1960s [72].
In Canada, single-phase switching was employed by BC Hydro for two 230 kV lines in
1964 and 1971. Similar to European cases, the reason for application was to increase the line
power capacity. This technology was used in Tennessee in 1970 for a 500 kV 150 km line [73].
Later in 1980s, single-phase switching was extended to 500 kV lines for transient stability
enhancement purposes [72].
2.3. Various Reclosing Methods 15
2.3 Various Reclosing Methods
There are various automatic reclosing systems in use today and they may be classified in dif-
ferent groups such as high-speed or time-delayed, single-shot or multiple-shot, and single-pole

or three-pole. These reclosing methods will be discussed in the following.
2.3.1 High-speed vs. time-delayed
High-speed reclosing means reclosing of the breaker which is not delayed, intentionally. The
only existing delay is the necessary time needed for deionization of the arc path.
However, most of reclosing schemes consider time-delay. In a time-delayed reclosure, the

breaker is being closed after an intentional time-delay which is longer than what is considered
for deionization of the arc path. The length of delay is mostly between one second and one
minute depending on the situation [5].
High-speed reclosing
There are some benefits in application of high-speed reclosure including reduction in outage
time, improvement in security and integrity of the system, higher probability of some recovery
from multiple contingency outages and preserving machine stability.
When high-speed reclosure is considered, there are also some limitation considerations on
system equipment, operating practices and configuration. Major limitations are as follows:
1) For permanent fault cases, stability of the system must be maintained after a high-speed
reclosing.
2) High-speed protection must be considered to trip at all terminals of the line, simultane-
ously for all fault cases to provide enough time for deionization of the arc path.
3) Configuration of the system has to be able to hold the angle difference of two sides the
opened breaker with a suitable variations.
4) Evaluation of any harm to rotational power devices connected to the system must be
considered.
5) Voltages induced from circuits in parallel must not sustain the faults.
For a successful high-speed reclosing, the path of the arc must be sufficiently deionized so
that the insulating properties can be re-established prior to re-energizing the transmission line.
There are some factors that are effective for a successful high-speed reclosing including:
- time of the fault clearance,
- construction and design of the transmission line,
- transmission line location compared to human-made or natural of any sources of danger,
- magnitude of the fault current,
- lightning stroke various components,
- weather condition,
- instantaneous voltage value at the moment of the line re-energization,
- inductive and capacitive coupling with the circuits in parallel connection,
- some other factors such as series capacitors, shunt reactors, tapped loads, etc.
Time-delayed reclosing
While time-delayed reclosure is applied, there are some factors to be considered including:
- Time-delayed reclosure can be beneficial provided that adequate protection system is not
applicable for all of a transmission line terminals.
- Decrement of the fault arc re-establishment probability by increment in the predetermined

time-delay.
- The desired minimum time for re-establishment of the system circuit, recognizing the
number of terminals of the line, stability studies, requirements of the system, kinds of the loads
and damping factors of the system.
Considering the time-delayed reclosure using the fastest reclosing schemes, usually only
one of the line terminals is being reclosed each time for re-energizing the line. Also, relays for
voltage monitoring are usually considered for determination of the system voltage presence on
one of the breaker sides. Therefore, the user will be able to select the conditions of the system
voltage that has to exist prior to issuing reclosing permission, in advance. Sometimes, users
allow a delayed reclosing only when the phase difference between the contacts of the open
breaker in a predefined range and stays synchronous with for a desired amount of time which
is performed by sync-checks.
Load-flow and stability studies must also be performed in case sync-check for unusually
long fault-clearance time is needed. Sometimes, double-checking of the phase difference and
being synchronous before application of the reclosure function is unnecessary due to the num-
ber of parallel paths with the faulted line.
2.3.2 Single-shot vs. multiple-shot
Single-shot reclosure means either high-speed or time-delayed reclosing the breaker during a
reclosing period, only for once. the other choice for reclosing, multiple-shot, is to reclose the
circuit breaker during the considered duty cycle for more than one time.
A typical reclosure at a line terminal may include a high-speed reclosing which is not su-
pervised by either sync-check or dead-line relaying, a time-delay reclosing which is supervised
by dead line voltage protection. A successful reclosing would allow the other line terminals
close by a time-delay, supervised by sync-check protection.
Generally speaking, modern high voltage and extra high voltage breakers are able to operate
in any multiple-shot reclosure period, practically. Some special issues regarding the circuit
breaker elements including opening or closing resistances, pressure of the available gas, air or
fluid, or factor of derating based on the expected fault current and number of fault interruptions
in a predefined time period in total can limit the effect on duty cycles or repetitive performances.
Before applying the multiple-shot reclosing, the user must make sure there would be no
danger of system instability or other possible limitations in case the fault is permanent. As
an example, multiple-shot reclosing is not practical when large motors or turbo-generators are
danger of mechanical damage is probable. Also, it is more desirable to postpone the second
reclosing shortly so that all transients of the previous reclosure are damped, properly.
Historically, single-shot reclosure of transmission lines has been in practice. But, later,
multiple-shot reclosure also became desirable and practical. This is mostly due to the fact
that most of the power system substations were not under supervisory control, traditionally.
In such substations, it is not practical by the personnel to perform the second reclosure after
one unsuccessful one. But, at substations where dispatching is performed by the personnel, an
unsuccessful single-shot reclosing may lead to a later outage which is not necessary.
However, at those locations where an employee must be dispatched, a delayed outage may
unnecessarily result if a single-shot reclosure is unsuccessful. About 25% and 10% of utilities
have records of successful 2nd and 3rd reclosing functions, respectively. But at the same time,
breaker constraints and reset time of the reclosing relays has to be taken into account not to have
problems in the result of additional operations of the breakers in a predefined time duration. If
there is no desire for single-shot reclosure schemes to be replace by multiple-shot methods, a

considerable success may be achieved by single-shot reclosing schemes just by increasing the
time-delay.
2.3.3 Single-phase vs. three-phase
The difference between single-phase and three-phase reclosing is that in single-phase reclos-
ing, tripping and reclosing is performed only to the faulted phase. In single-phase reclosing,
electrical power can be transfered through the non-faulted phases while the faulted phase is
open. This leads to decrement is rotor-angle drift rate between the synchronous machines and
increases the stability margin of the system. The only major problem is affection of the ground
protection of the system due to circulation of the ground currents.
During fault clearance in a single-phase reclosing, due to capacitive an also inductive cou-
pling of the phases, there is an induced voltage in the de-energized line. The magnitude of this
voltage depends on the mentioned capacitance values and therefore, on the transmission line’s
configuration and physical dimensions. This voltage may causes the arc to stay on for a longer
time period while the faulted line is isolated. In this case, the arc is called the secondary arc.
The main purpose of considering the reclosing time-delay is to give enough time to the sec-
ondary arc to be extinguished and therefore, to avoid reclosing-onto-fault and arc restriking.
The amount of this time-delay, called the dead-time, is chosen based on the system analysis.
One of the differences between single-phase and three-phase reclosing schemes is the differ-
ence in the dead-time which is considerably shorter in three-phase reclosing.
In some cases of single-phase reclosing, the amount of the induced voltage on the secondary
arc is really huge which can make the arc extinction time very long or even infinity. In such
cases, one practical method for decrement of the arc extinction time is to compensate the line
capacitances and therefore, to reduce the amount of the induced voltage. This is basically
done using for-legged shunt reactor, i.e., a three-phase reactor with a neutral reactor. Another
possible method is to close and reopen the grounding switches installed on the faulted phase in
a fast manner.
In some special power system substation designing methodologies, single-phase tripping/re-

closing and three-phase tripping/reclosing are considered independent and non-related. In such
designs, it is the role of the protection system ro recognize the fault type, i.e., single-phase or
multi-phase and also to introduce the faulted phase(s). In general, three-phase tripping (with
no reclosing) are applicable when:
- A single-phase-to-ground fault is followed by an unsuccessful reclosure,
- There is a multi-phase fault,
- There is an abnormal delay in performance of single-phase reclosure task while the fault
is recognized as single-phase-to-ground,
- More phases become involved in the fault while the main faulted phase is isolated.
2.4 Protection System Operation During Reclosing
When the reclosing task is considered for the protection system, the sequence of operations can
be listed as 1- single-phase isolation of the faulted phase and 2- reclosing of the open phase.
If the reclosing process is successful, it means the isolated phase is back to operation, success-
fully, and the normal performance of the power system is expected afterwards. However, if
the reclosing is not successful, the next action is either restarting the process by isolating the
faulted phase for the second time or refusing the reclosing for the second time by three-phase
tripping of the breakers.
Sequence of operation of the protection system during a successful reclosing is shown in

Figure 2.1 [74]. As obsrved, there is a delay for operation after the fault instant, called oper-
ating time which represents the fault detection delay. Afterwards, another delay is considered,
intentionally, called dead time. This time duration consists of 1- opening time, i.e., the time
needed for the contacts of faulted phase to separate, 2- arcing time in which the secondary arc
fed by the charge of the line and also by the non-faulted phases becomes extinguished and 3-
waiting time for reclosing. After the dead time, reclosing command is issued and after a delay
which is for the closing time of contacts, the system becomes energized and goes to normal
operating mode. The typical value for the dead time setting varies between half a second to
one second [3].
2.5. Application Conditions 21
Figure 2.1: Sequence of operation of the protective relay during a successful reclosing
[74].
Reclosing can be unsuccessful due to permanent fault inception of reclosing before the
secondary arc extinction. If the reclosing is unsuccessful, the sequence of operations is more
complicated. As shown in Figure 2.2, the sequence of an unsuccessful reclosing is similar

to the successful one until the first reclosing. Then, as the fault is not cleared yet, the whole
procedure may become reset by starting from the beginning and energizing the trip coil and
trying the whole process for couple of times. If any of the reclosing trials are successful, the
system can go back to normal operation. However, the number of trials is limited and finally,
the last action will be to issue the three-phase trip signal and to de-energizing the whole line.
2.5 Application Conditions
Lines with transformers and series/shunt compensation
In practice, necessary inspections and repairs must be performed for a faulted transformer
and it is not a common practice to bring it back to service, immediately. Therefore, the line
Figure 2.2: Sequence of operation of the protective relay during an unsuccessful reclosing
[74].
breakers connected directly to a power transformer are not permitted to perform reclosing in
case there is a fault inside the transformer. Therefore, by operation of the relays associated
with the power transformer, reclosing task has to be blocked. In such cases, the pre-considered
lockout relay will be initiated by the transformer relay and the transfer trip signal will be
sent directly to the the part considered for reclosure tripping/blocking. In such a process, it is
important to get the necessary permissions for terminal breakers’ reclosure when fault happens.
For shunt compensated transmission lines, the same points mentioned for transformer-
connected lines must be taken into account. In case a shunt reactor fault is detected, it is
essential that the line breaker be tripped, remotely, by initiation of direct transfer trip signal. It
is generally desired that these shunt rectors stay connected to line while being connected to the
grid. In such cases, it is very important that the reclosing be prevented if any fault is detected
in the reactor. The only practical solution is to isolate the faulted reactor in advance, provided
that the line operation with no reactor is permitted by design of the system.
Another problem that may be arised is the voltage oscillations produced in the transmission
lines due to the resonance between the line capacitances and the shunt reactors, following
the line isolation. The problem is even more sever if the oscillation frequency reaches the
frequency of the power system which will need much more time-delay for damping. Therefore,
immediate line re-energization by high-speed reclosure with no delay consideration would not
be applicable. At the other hand, the oscillations may cause errors in CVT output and also
mall-operation of the relays, specially frequency and voltage.
When there is a fault in the line, series capacitors could be bypassed if they are controlled,
automatically. Afterwards, the capacitor bank can be re-inserted, automatically, if the reclo-
sure is performed, successfully. In such cases, reclosure is applicable as switching transients

are mostly considered in design of the system.
Transmission cables and gas-insulated buss
For transmission cables, auto-reclosing usually is not applicable. one exception is when
the line is a combination of cable and overhead lines, provided that the overhead part is the
major part. Sometimes, even the overhead line and cable have different protection systems and
therefore, it is possible not to perform reclosing when the cable is faulted.
When dealing with gas-insulated buses, auto-reclosing is rarely applicable unless it is par-
tially involved in the line relaying, e.g., part of the bus between breaker-and-a-half method
breakers. However, the gas-insulated bus portion could have its own relaying system. In such
conditions, the bus fault reclosure blocking will be possible.
Multi-terminal/breaker lines
Regarding multi-terminal transmission lines, assuming that for all terminals, the stability
of the system is not disturbed by reclosing on permanent fault and high-speed reclosing is
considered for all terminals, It is possible to apply high-speed reclosing. Generally speaking,
it is easier to apply high-speed reclosing to a two-terminal line than a multi-terminal one as at
least a shorter dead-time will be required.
When applying reclosing to a multi-terminal line, the pilot relaying channel has to be con-
sidered. Also, relays must be set for each of the terminals to watch for all possible faults at that
spot. This leads to more accurate performance as all terminals are being taken care all together.
However, it could be more interesting to perform high-speed reclosing at one of the terminals
and monitor other terminals for permanent fault detection.
Sometimes, it is practical to connect the line ends to more than one circuit breakers when
dealing with extra high voltage systems. Breaker-and-a-half, double-breaker and ring-bus con-
figurations are some examples for this purpose. For the fault clearance in such circuit arrange-
ments, more than a breaker needs to be tripped.
However, applying the reclosure task to all breakers at the same time can lead to more
fault duties. This means, the preference is to perform the line testing by applying reclosure
to just one of the breakers and then move to the rest in case of success. This can be double-
checked using sync-check and also by monitoring of the voltage. Sometime, the first reclosing
is considered as high-speed.
Transformers
There have been vast studies performed on power transformer short circuit testing and even
there were efforts to modify the standard test circuit to deal with the increasing rate of the power
transformer failures due to the fault currents, specially for old transformers manufactured in
1960s. The major damages that fault currents arise in transformers can be mechanical and
thermal which are interconnected. The mechanical stability of a transformer depends on factors
such as the size of the clamping forces and the quality/dimensions of the materials used for
manufacturing. It also depends on the instantaneous temperature of the transformer which is
in a correlation with the density of the winding currents.
There are also some studies performed on transformer loading. If the currents in the wind-
ings are symmetrical, there would be less amount of stress compared to the asymmetrical cases.
During a fault, different components of the transformer are under different stresses and ten-
sions, and therefore, deformations of different types are expected for different parts, temporar-
ily or permanently. In case of a permanent deformations, it means the amount of clamping

forces have not been adequate and the transformer is subjected to failure due to future faults.
In the literature, short circuit studies are being performed mostly for the cases of small
transformers and in distribution level. There are also similar problems for power transformers
in transmission level. However, there are more complexities in studying the effect of reclosing
on power/transmission transformers due to number of windings and components involved.
High-speed reclosing also affects the thermal capabilities of power transformers during
short circuit. Although, thermal capability is not a very critical factor, it is essential to be stud-
ied when high-speed reclosing is considered. The last factor that can affect the transformer life
is the loading habits. In some cases, power transformers are subjected to various overloading
conditions that can decrease the short circuit capability of transformers and therefore, some
restrictions in performing the reclosure must be considered.
Generators
There could be considerable amount of stress in shafts and other parts of turbo-generators
because of operation of switches which has been of research interest, recently. However, there
are not that much documented reports showing that harms or damages are actually caused by
switching or reclosing operations. But the fact is that, such stresses, even during an ordinary
switching operation, have accumulative effects which means simple operations such as reclos-
ing can damage the electrical machines by time even though not be the main reason.
Current and power oscillations/transients caused by opening/closing breaker operations can

lead to damage and stress on the generator units close to that point. This will affect many turbo-
generator components including stator components of the generator such as winding and core,
all the rotary elements in the turbine, generator and exciter, and shaft. The main problem is
when for a closed breaker, the average power changes. This leads to torsional stress production,
specially in the rotary components of the turbo-generator
To prevent such damages, some limitations must be applied to rate of change of current and
power depending on the power factor and the rated power, to be considered just for steady-state
conditions. For damping of the oscillations, a delay as long as ten seconds seems adequate and
by this assumption, a switching operation can be applied.
Another problem is what happens after clearance of a close-in fault close to generator buses
if high-speed reclosure is considered. In such cases, the useful life of rotary components will
be used which leads to the machine failure. If high-speed reclosing is considered, for any
multi-phase fault cases, the level of stress will be less than thirty faults (which is the worst
case scenario) but still the fault magnitude is enough for causing the life losses. High-speed
reclosing onto single-phase-to-ground fault may provide less stress than multi-phase faults.
Although, the chance of failure is not high for such a fault type, the suggestion is to do studies in
any case, including high-speed reclosing effects on the geographically far but electrically near
substations as huge amounts of stress might be applied to some possible remote generators.
The main studies for turbo-generator shaft failures caused by torsional oscillations is not
easy at all. Such studies may include shaft and rotary system modeling, obtaining fatigue pa-
rameters from the torque time response and finally, extracting a considerable screening frame.
Any of these steps could be a considerable topic for studying. However, the most essential part
now might be the last part, i.e., to find some practical measures to be used in the process of
design of power system substations. There are also some research works indicating that it is
not feasible to study the shaft fatigue using very ordinary measurements like the rate of change
of the current or the active power.
Considering all the mentioned parameters above, utilities have chosen very different strate-
gies in different conditions. In some cases, they have ignored application of high-speed re-
closing due to the possible harms, specially for power plant substations and instead, have em-
ployed some alternative reclosing schemes such as considering time-delayed reclosing, using
sync-check and even not to perform reclosing. One other possible practice has been to apply
time-delayed single-phase reclosing instead of high-speed three-phase reclosing.
In case auto-reclosing is not considered, the line will be tripped off and therefore, the par-
allel circuits will have to carry more loads which can lead to the over loading relay operation.
Therefore, it might sometimes be better to accept the risk of the machine failure not to allow
the phase angle to increase. In such cases, the reclosure method and the protection relay must
be chosen, properly, to decrease the amount of the stress on the rotary parts of turbo-generators.
This also needs more precisely selection of the relay/recloser settings as the situation can be
extremely harmful and stressful.
Motors
Basically, there are considerably large similarities between generators and motors. There-
fore, it is expected that most of the items mentioned for generators be valid for motors as well.
Considering the rotary elements and the shaft, stresses similar to generators exist in induction
and synchronous motors too. Motors in power systems are subjected to big stresses and dam-
ages caused by transients and other electric disturbances, and the amount of stress increases by
increment of the motor and the associated drive systems. Switching operations in the power
system can also cause such stresses on the drive systems by providing transient torques.
Also, similar to the generator issues, the level of stresses mostly depends on the system
structure and configuration. The connected transformer impedances will act like buffers and
will absorb the disturbances caused by switchings, considerably. When synchronizing with the
power system, it is generally easier to perform the synchronization process for motors.
Another point to consider for isolation of disturbances is the protection speed. In fact, for
making the reclosing more effective, it is very essential to use fast protection systems to clear
the fault as fast as possible. During the reclosing time-delay, called the dead-time, it is also
possible to isolate the motor form the rest of the grid, temporarily. In such cases, there is not
that much limitations in the number of high-speed reclosing attempts and reclosing rates as
large as twenty to thirty are also considerable. In such conditions, it is possible to reclose the
breaker of a source, however, it is more desire to predict fast protective devices for the voltage,
angle and frequency. Based on the connection of the primary side, it is better to separate the
load from the motor by issuing the trip signal or by giving permission to the breakers to operate
within a limited angle and voltage range.
Generally speaking, it is not easy to talk about the angle/voltage limits to be considered
for re-energization of isolated motors, either automatic or manual. But, there had been some
practical values in 1930’s for both angle and voltage as 6◦ and 25%, respectively. Also in
transmission lines, the rate of high-speed reclosing is considerably low and therefore, it will
not affect motor loads and fast transfer when there is no supervisory device. In addition, it is
fast enough not to allow the residual voltage decay reach toleration limits.
Regarding the design/application personnel, they may consider the factors mentioned above
as each of them can cause permanent damages through transient torques, for example. To avoid
this danger for motors, it is necessary to study these aspects of the system, precisely.
Larger motors, both induction and synchronous, can potentially cause problems in reclos-
ing. But, in many processes in industry, when there are a huge number of small machines
working at the same/close spot, high-speed reclosing can help improving the service offered by
utilities. Therefore, they may advice the consumers about merits of this trick and also, in de-
sign it can be considered. Electrical energy providers and consumers both may know about the
possible system transients during a switching process. In fact, the risk evaluation and possible
solutions can be considered only if both sides are fully aware of the situation, in details.
2.6. IEEE Guides for Auto-Reclosing 29
2.6 IEEE Guides for Auto-Reclosing
2.6.1 Fundamentals and applications
In this subsection, basic ideas of auto-reclosing are described for distribution and transmission
system circuit breakers. Usually, the preference is to reclose the breaker after the line protection
system tripping operation. In case the fault is not permanent, the isolated line can be brought
back to operation and therefore, the system will be able to carry full active/reactive power
again, and the obvious results will be increment in the system stability margins as well as
offering a more reliable service to the consumers. The same scenario is probable if there is a
fault on the substation bus at which reclosing can help to restore the bus in case of a temporary
fault.
For having a proper auto-reclosing, consideration of some items is essential including:
1) How successful is the reclosing?

2) How much is harmful to reclose onto a fault?
3) Is the any need for specialized interlocker to avoid reclosing under special conditions?
4) High-speed or time-delayed reclosing?
5) In case reclosing was not successful, how many attempts should be considered?
6) Is there any need for voltage supervision?
7) Is there any need for sync-check?

8) In what conditions, reclosing must be avoided?
9) In multi-terminal lines, which terminal must be chosen for the first reclosing attempt?
10) How to initiate auto-reclosing?
Auto-reclosing task performed by a reclosure relay
Historically, auto-reclosing task has been done by reclosing relays which were reasonable
for some special circuit breakers. But, it is possible to perform the reclosing task by any number
of processes and devices which has an input to get if a breaker is closed or opened and even
how the tripping signal is applied, either by a relay or manually.
Generally speaking, it is possible to perform auto-reclosing task in one of the following
conditions:
1) Reset condition: In this condition, the breaker is closed at first. Then, the auto-reclosing
task waits for receiving a signal for starting the timing for reclosing following the breaker’s
opening.
2) Cycle condition: In this condition, the breaker is open and the auto-reclosure task is
timing to the set time for closing the breaker. The other case is when the breaker is closed
after a successful auto-reclosure, timing to the set time for the purpose of returning to the reset
condition.
3) Lockout condition: In such a condition, the auto-reclosure is in the close position and is
locked in that condition. The reason is reclosing has happened at least once before, however,
the relay has issued the trip signal again. In this condition, the breaker will stay at locked
condition and the reclosing task can not be performed any more. The only way to reclose
the breaker in this condition is manual operation, either locally or remotely (local-manual or
remote-manual).
4) Power-up condition: In this condition, the associated relay reads the input signals and
the breaker situation to determine the true condition and then will move to that condition. For
example, in normal conditions when no special input signal is initiated, if the breaker is open,
it will move to the lockout condition, and when it is closed, reset is the true condition.
If the auto-reclosure task is in the reset condition, this task may be performed in two pos-
sible ways for starting the timing to the first attempt which is based on the logic scheme or the
design of the device:
- To detect the breaker’s opening, which will start the first reclose/open interval timing,
or
- To Detect that the breaker has been tripped by the protection device, and to confirm the
breaker opening afterwards, which will start the first reclose/open interval timing.
During opening of a breaker which is controlled, the reclosing protection device will start
timing to the first auto-reclosure attempt. Such a transition from timing to the initial auto-
reclosure attempt will change the reclosing task from the reset condition to the cycle condition.
While completing this delay, the reclosing protective device will close the output contact to
initiate breaker closing and will confirm that the breaker is been closed due to a condition
change of the input of the breaker status. If the breaker could not perform the closing task
during a preset time, the reclosure protection device might move to lockout condition, continue
to assert the close output or release the close output, based on the design of the reclosure
protection device and program.
During the closing of the breaker, the reclosure protective device will start the reset timer
which is called reclaim timer too. When breaker stays closed, means a temporary fault was
detected, the protective device will continue timing to reset. When the reset time-delay is
completed, the auto-reclosure task will return to the reset condition and will remain there for
the time the breaker stays closed.
For the cases of no fault clearance and the protection method shows the second tripping
signal, the reset timer is stopped and reset, and the reclosure protective device starts timing to a
second auto-reclosure attempt provided that the relay is set or programmed for more than one
auto-reclosure attempt.
Following any of the attempts for auto-reclosing, in case the breaker stays closed for a time
as long as the timer of the reset delay, the reclosure relay will return to the reset condition.
For the permanent fault cases, if number of the reset auto-reclosure attempts finishes, the
next condition of the relay would be lockout while the breaker will stay open. This condition of
the reclosure task will be kept until the breaker becomes closed by other methods, e.g., manual.
Based on it’s design, the protective device might reset right away of after a time-delay. The
amount of the delay can be equal to the successful reclosing reset delay or can be different.
For permanent fault cases, if the breaker is closed, manually, but becomes open again,
the manual reclosing must be blocked. Similarly, it will be reasonable to avoid the reclosure
device from resetting provided that the breaker closure leads to the protection system pick up
and starts timing to trip. In case the tripping time-delay be larger than the reset time-delay,
a pumping condition of the breaker may happen which finally can cause the damage for the
system equipment and breaker, and also some other hazards will be possible.
2.6.2 Auto-reclosing for transmission systems
Transmission systems overview
Loosing a transmission line in the power system will have a considerable effect on the sys-
tem’s reliability and economics. Auto-reclosure is been performed in a line for minimizing this
effect. Time-delayed and high-speed reclosing methods are effectively and widely employed.
The access to a huge amount of information using intelligent relaying system and SCADA
will permit more selective reclosure tasks which can makes the reclosing a more successful
operation at one hand and can reduce the rate of damage to the power system due to reclosing
onto permanent fault at the other hand. By increasing rate of adaptive reclosing techniques,
controlled breaker closing and single-phase trip application, the reliability of the system has
improved and the negative effects of auto-reclosure failure has decreased on the system.
Additionally, beside the auto-reclosure methods, there are more system elements that need
to be considered regarding their impacts on the reclosure application. As an example, in a
line which needs reclosing supervision by sync. machine. For the system effect reduction,
reclosing can be applied just to one of the line ends. There are other elements that may affect
the reclosure are transformers, motors, generators, reactors, capacitors, multiple-terminal lines,
etc. The auto-reclosing methods and the system element that may affect the reclosing of the
breakers in a transmission line are discussed in the following in details.
2.6.3 Auto-reclosing methods
High-speed auto-reclosing
High-speed auto-reclosure scheme means to automatically close a circuit breaker with no

predetermined delay more than what is adequate for deionization of the arc (see IEEE PSRC
Report [2]). In order to have successful operations at auto-reclosing attempts for temporary
faults, it is suggested to consider high-speed auto-reclosure schemes.
There are some benefits in application of high-speed reclosure including reduction in outage
time, improvement in security and integrity of the system, higher probability of some recovery
from multiple contingency outages and preserving machine stability.
When high-speed reclosure is considered, there are also some limitation considerations on
system equipment, operating practices and configuration. Major limitations are as follows:
- Configuration of the system has to be able to hold the angle difference of two sides the
opened breaker with a suitable variations.
- Voltages induced from circuits in parallel must not sustain the faults.
- For permanent faults, the system must be stable after a high-speed reclosing.
- High-speed protection must be considered to trip at all terminals of the line, simultane-
ously for all fault cases to provide enough time for deionization of the arc path.
- Evaluation of any harm to rotational power devices which are connected to the power
system must be taken into account.
For a successful high-speed reclosing, the path of the arc must be sufficiently deionized so
that the insulating properties can be re-established prior to re-energizing the transmission line.
There are some factors that are effective for a successful high-speed reclosing including:
- construction and design of the transmission line,
- time of the fault clearance,
- magnitude of the fault current,
- transmission line location compared to human-made or natural of any sources of danger,
- weather condition,
- lightning stroke various components,
- inductive and capacitive coupling with the circuits in parallel connection,
- instantaneous voltage value at the moment of the line re-energization,
- some other factors such as series capacitors, shunt reactors, tapped loads, etc.
2.6.4 Auto-reclosing settings
Number of reclosing attempts

The simplest auto-reclosure task in terms of number of attempts is single-shot reclosure,
i.e., to reclose the breaker only once. Single-shot reclosure can be performed either delayed or
high-speed.
The other option is to perform reclosing more than one time which is called multiple-shot
reclosure. For performing multiple-shot reclosing, many factors such as the system stability
considerations, possible system equipment failure due to fault currents, possible damages to
the customer and other effects, etc., must be taken into account.
Dead-time
There are some items to be considered for reclosure attempting when the relay has issued
the trip signal. It is necessary to bring the faulted phase back to operation as fast as possible.
However, if it is performed too early, considering a temporary fault case, there is a chance
that the associated arc is not extinguished yet, and therefore the reclosure attempt will be
unsuccessful as the arc will restrike.
During an arc restriking, the arc path, which is not deionized yet, will reconduct after an
unsuccessful reclosure. The time interval which is needed for deionization is a function of
the spacing of the conductors, the voltage level, the weather condition and the fault current
magnitude [81]. A practical value for a minimum needed time-delay to avoid arc restriking can
be obtained from (2.1) [5].
t = 10.5 + VL /34.5 (2.1)
in which t is the time-delay, in power system cycles, and VL is the line rated voltage, in kV.
For some circuit breaker types, there is a need for adding a delay to either the auto-reclosure
time or the breaker time to reach the deionization time of the arc. As an example, there are
breakers available capable of reclosing in two to three cycles. For using these breakers, it is
required to add a delay to prevent from reclosing while the arc is not extinguished yet.
The deionization time of the arc may increase even more in case single-phase tripping is
employed and in circuit phases keep the arc on, or when a side parallel line supports the arc to
be sustained.
Auto-reclosure reset-time/lockout
After a successful auto-reclosure, the protective device must be reset and become ready for
the next possible fault inception operation. This is done by a timer called reset timer. In case
of a successful auto-reclosure, the reclosing protective device will go back to the rest condition
after the reset time-delay. The reset time-delay selection depends on the fault nature and the
expected fault clearance time. As an example, the chance of re-tripping and therefore, another
reclosure will decrease if a longer reset time-delay is chosen. If the delay is shorter than the
time needed for the arc extinction, then another tripping/recloser might happen which has its
own effects on the system and the consumers. At the other hand, if the reset time is too long,
some lockout operations might happen although excessive reclosing task will be avoided.
In case the auto-reclosing operation is unsuccessful, the protective device will move to
the lockout condition in order to avoid the interruption device from automatic closing. Auto-
reclosure is either locked out or terminated afterwards the predetermined attempts to the line
re-energization are tried and completed, unsuccessfully. The lockout state can be beneficial to
prevent too many wear on the relay from multiple operations caused by common line temporary
faults, such as tree branches broken by wind, etc.
There are auto-reclosure systems which are capable of being programmed to prevent se-
quential auto-reclosure operations by providing a lockout situation in case the fault number
reaches a predetermined value in a time window. For example, a 3-shot auto-reclosure method
can be programmed to lockout in case 7 faulted cases occur in a 30-minute time window.
Another benefit of lockout is to prevent a relay from unwanted auto-reclosure while the
closed condition is applied to the line, manually. This condition will be applied, automatically,
if the breaker is manually closed, either remotely or locally, and the breaker will remain closed
for a predetermined time period. Therefore, no fault will exist at the breaker closing time.
2.6.5 Auto-reclosing blocking
Auto-reclosure task is blocked typically in one of thee following conditions:
1) Operator tripping: Auto-reclosure will be blocked in case the breaker is opened either
by remote control or manually, at the station . When this kind of opening happens to a breaker,
the preference is to be under the control of operator for for closing.
2) Voltage supervisory monitoring of the line: Auto-reclosure could be blocked when there
is voltage on the line. This kind of supervisory monitoring is considered usually when the line
is connected to large generators, motors or other solid sources. Auto-reclosure will be blocked
when these sources from downstream maintain the voltage on the line to avoid probable dam-
ages to the related rotary devices. The reason for the failure is establishment of an unwanted
system operation condition or out-of-phase energization.
3) Fault in an underground cable, a transformer or a bus: This kind of fault is naturally

permanent most of the times and a breaker’s auto-reclosure can deepen the damage to that
equipment. Therefore, auto-reclosing’s both benefits and risks vs. blocking the auto-reclosure
must be considered.
4) Unbalance of voltage: Auto-reclosing may be blocked when an unbalance is detected in
the substation voltage. An open phase in the source-side may lead to such unbalances. Service
restoration when the condition is not balanced can lead to customer equipment damage.
5) Circuit breaker failure protection: Circuit breaker failure protection basically trips all
breakers which are directly connected to the bus or connected to the bus through the differential
protection of the bus. Auto-reclosure of the breakers will be blocked till isolation of the failed
breaker and restoration of the bus.
6) If closing of breaker fails: In case a breaker fails to close following an auto-reclosing

attempt (auxiliary contact 52), closing take more time than expected, or detection of an open
circuit (discontinuity), auto-reclosure attempts will be blocked to avoid further damages.
7) High-current faults: For blocking of auto-reclosure for close-in high-current faults, a

high-set instantaneous element may be employed. This blocking type is usually applicable
when the fault reaches the rating of damage of the generator transformer or other equipment.
The fault type is probably permanent in the exit cables or equipment of the substation.
8) Cumulative operations lockout: This case is considered where the duty of fault ap-
proaches the preset rating of circuit breaker for auto-reclosure blocking after a predefined op-
eration number till performing the necessary maintenance and inspections.
9) Receiving a transfer trip: Sometimes, a timer is initiated when a transfer trip signal is
received. The reclosing sequence operation is avoided in case of the timer’s time-out.
10) Backup tripping: A zone 3 distance zone (overreaching) employed for backup protec-
tion can be used for reclosure disabling as it shows a protection failure of another zone.
11) Under-voltage or under-frequency load shedding: Auto-reclosure is basically blocked

for feeders tripped by load shedding.
2.6.6 Operation considerations
Lines with underground cables
For underground cables, faults are mostly of permanent nature and auto-reclosing must be
applied according to the following guidelines. Lines that use underground cables partially and
not for the whole length show special concerns for utilities to consider auto-reclosure.
If the line is made of cable, completely, auto-reclosing must be avoided. It is not needed
to put the cable at risk of more damages and subject the power transformer of the substation,
circuit breakers, buses and other parts of the power system to more potential damages and
stresses. According to an IEEE Power System Relaying Committee Report, auto-reclosure ap-
plication for underground cable lines is not recommended [2].
Based on the length and location of the cable portion compared to the total length of the line,
utilities sometimes consider this feeder type as completely overhead and apply their own auto-
reclosing practices. Some other utilities may consider a single-shot auto-reclosure to evaluate
the line for a temporary fault or other possible conditions that might lead to misoperation of
the relay. The other solution can be to use a sectionalizer for isolation of the cable part and a
modified version of the auto-reclosure schemes, simultaneously. If the line is partially overhead
and partially cable, auto-reclosing can be applied if absence of fault in the cable portion can be
determined [2].
It must also be considered that which end of a combined line is to close-in as the first auto-
reclosure attempt. Overvoltage studies based on a transient network analyzer might need to
be performed to check for any possible problems may happen in case auto-reclosing is applied
to the line. Such studies could be extremely essential in case a shunt reactor is connected to
the line for controlling the effect of capacitances of the cable which increase the line voltage,
called Ferranti effect [1, 2].
Auto-reclosing following bus faults
Auto-reclosure task following a fault on bus is sometimes applied to return the system to
normal condition. The decision to use auto-reclosure task following a fault on a bus is made
as a result of a trade-off between the reclosing consequences into a bus fault and the effect of
an extended bus outage. As the buses of the power system usually end in couple of elements,
both the bus fault and the outage of the bus might have considerable effects on the operability
and the security of the system. The merits of auto-restoration of a bus after a temporary fault
are usually evaluated versus the possible risks if the fault in fact is permanent. Some possible
risks can be mentioned as probable damages to basic equipment, specially turbo-generators

and transformers, and impacting the stability of the system.
For substations of type of open-air, in some cases, auto-reclosure on the breaker of a bus is
done when protection device operates for the bus faults. Generally the auto-reclosure is delayed
in time, e.g., for five seconds, and there is not supervision unless if buses could be connected
to other sources. Such sources can be distributed on the transmission line or feed other trans-
formers. An under-voltage relay might supervise the auto-reclosing in such conditions. Other
conditions of the system may need for the auto-reclosing blocking, e.g., under-frequency, trans-
former differential, breaker failure, etc.
Similar to overhead distribution or transmission lines, a bus is subjected to faults external

causes such as insulator flash-over, animals, lightning, etc, when employs open-air primarily
as the insulation medium. An open strain bus or outdoor open rigid bus are examples of this
bus type. If these fault types are cleared, immediately, a successful reclosure with no need for
inspection would be probable. This probability may be affected by spacing of the conductors
which itself depends on the voltage level of the bus. Temporary faults incepted due to presence
of an animal are more likely to happen in distribution networks.
Open-air buses are not usually that exposed to public due to presence of substation barriers,
fences, etc. However, this availability is lower for transmission lines. Therefore, application
of reclosing in open-air buses could be considered as a greater threat for public that transmis-
sion line reclosure. Another danger of reclosing of open-air buses is for substation personnel
during maintenance and other routine tasks. This second danger can be decreased by using
the function of reclose cutoff which blocks the auto-reclosing during routine tasks, etc., in the
substation. This generally will be similar to the live-line maintenance function which is usually
considered for transmission and distribution circuits.
It is possible to protect a bus from external interference, physically, by an appropriate insu-
lation method. Isolated phase bus, SF6 switchgear, metal-clad switchgear and insulated cable
are examples of this kind of isolation. If a fault happens ion such buses, it will not be that likely
to be cleared by itself and will not be beneficial to apply auto-reclosing to such kind of buses.
When auto-reclosure scheme is applied to a bus fault, it is possible to consider single-shot

reclosure protective device for one of the breakers connected to the bus, to perform the breaker
reclosing in case of dead-bus-live-live condition. In practice, it is common to choose a breaker
for performing the testing task for the bus that will have the lowest effect on the system in case
of permanent fault. Most of the times, it is the live circuit which has the lowest short-circuit
capacity. But, the differential relay of the bus must be sensitive to the downgraded fault current
level so that for permanent fault cases, tripping becomes possible. One solution for this is to
decrease the pickup current threshold for a short time period during de-energization of the bus.
Another solution can be considering an instantaneous over-current relay dedicated to reclosing.
For the purpose of choosing a proper level of sensitivity for reclosing protection of the but, the
effect of inrush currents provided by VTs and other bus equipment must be taken into account.
While reclosing the breakers connected to one bus, if the reclosing operation was success-
ful for the first bus, auto-reclosing can be performed for the rest of the breakers of the bus
considering the limitations and permissions issued for the live-bus condition.
Circuit breakers
A minimum amount of time between initiation of tripping and reclosing signals is needed
for all interrupters and circuit breaker mechanisms. Such a gap in time is essential for the
breaker insulator to stabilize the internal insulator, otherwise, the breaker will not be able to
tolerate the rated electric field and will collapse in case the breaker reopens the circuit due to
presence of fault. If this time delay is not included in the design of the breaker, it must be
considered by the utilities and protection engineers. The minimum needed time is defined by
the manufacturer of the circuit breaker.
The other factor to be considered regarding circuit breakers is the time needed for recharg-
ing the springs, compressed air and other mechanical components of the breaker for the next
operation. The subsequent operation can not be performed, properly, if the recharging time is
too short. In such conditions, an interlock scheme might be employed by the breaker so that
adequate amount of energy is restored for the next operation.
Such features are mentioned in the associated standards for air and oil breakers’ magnetic
circuits in which the time gap between consequent faults is too short for strength recovery of
the dielectrics. Such de-rating factors might not be needed for vacuum or SF6 puffer circuit
breakers because for these technologies, the recovery time of the dielectric is shorter than
open-interval time. But, if the duty cycle of the circuit breaker is different from the standard
requirements, it is recommended to consult the manufacturer about the issue. For the case of
oil circuit breakers, the capability for the circuit interruption is dependent on the fault current
magnitude and the sequence of reclosing.
It is essential and required for circuit breakers to have a capability of interruption based on
the standard duty cycle, Rated Standard Operation Duty. According to IEEE Std C37.04, the
standard duty cycle is defined as two operations with a 15 s of interval between them or in other
words, CO + 15s + CO, where CO means close-open. If the time interval between operations
is longer than 15 s or the setting of duty cycle characteristic of a circuit breaker is performed
for more than two operations, the interrupting capability rating of the breaker will be modified.
One or two of these characteristics can efficiently be incorporated by the auto-reclosing duty
cycle to derate or reduce the interrupting rating. For allowing the recovery of the dielectric of
the circuit breaker insulations, such derating is required.
Lines with automatic sectionalizing
For decreasing the amount of loads affected by short circuit fault, auto-sectionalization
option might be used by radial lines which are connected at both ends to source buses, and
distribution feeders. Also, for isolation of the faulted section of the line and therefore, the auto-
reclosure permission for service restoration for the non-faulted sections of the line, automatic
line sectionalizers or similar methods might be employed by radial distribution circuits.
Automatic sectionalizing might also be used in transmission lines with networked structure
and tapped loads for improving the quality of the service to the loads. In this case, on each
end of the line a potential device is required as an ideal practice to allow auto-reclosure voltage
supervision. But, it is possible to coordinate the line-source breaker’s auto-reclosing task with
sectionalizing method to allow sectionalizing following the first shot and before the second
shot of the auto-reclosing. It is better to perform the first shot of the auto-reclosing prior to
sectionalizing. By this, it will be possible to restore all parts of the transmission line right
without any delay. For this, the fault has to be temporary. In case of failure of the second shot,
the line will be locked out as there would be an indicated fault between the sectionalizing spot
and the source.
There is a similar reclosing considerations for networked or radial lines with loads under
application of auto-transfer methods. Such methods are employed for transferring to reserve
or second supply connected to another feeder or line. Timer of transferring will be initiated
as a result of the primary supply loss. The transfer will be aborted and the transfer timer
will stop and reset if the fault of the line is temporary and the primary supply is re-energized,
successfully, before the time out of the transfer timer. Therefore, for coordination with the
transfer time delay, the first reclosure attempt must be made with a short time delay. The sec-
ond attempt should not be made till the auto-transfer method has time to perform and finish the
transfer function in case the first reclosure attempt is not successful.
Adaptive auto-reclosing
Nowadays, it is possible to implement many reclosure methods that can be adapted to dif-
ferent conditions including time of the day, heavy load or weather forecast. Such methods
mostly depend on both the final consumer and the system. Following are some examples of
installed and proposed adaptive reclosure methods:
1) A method disables or enables fuse saving by performing multiple-shot reclosure during

weekends and evenings. The worst case scenario for an industrial customer is to experience
any interruptions in the facility operation. Therefore, during operation hours it is not allowed
to do fuse saving.
2) Operators are allowed by the method to disable or enable auto-reclosure depending on
the weather conditions. There are some historical data provided by utilities showing that most
of successful auto-reclosure operations have performed during thunderstorms.
3) Controlling the sequence of auto-reclosure on a line including downstream reclosers,

using overcurrent element. In such method, an overcurrent relay contact is being closed when
the current level reaches to a value associated with the downstream fault current. In case this
contact is closed, it can be concluded that the downstream recloser has operated. This leads to
the reclosing relay skip the initial reclosure attempt if ultimately the breaker is called on for
the fault interruption.
4) Application of overcurrent element for detection of the load level to show if a consumer
has started a large motor and it is desired that the recloser does not operate for such a condition.
5) Application of an impedance element or an overcurrent element of high-set for determi-
nation of any fault inception in the cable portion, between the transition to overhead and the
breaker. For such a case, reclosure must be avoided.
6) Alarm of trip circuit monitor or breaker failure trip. In case a relay issues the tripping
signal while the breaker does not open for during the time it should, for example 6-10 cycles,
auto-reclosure is better to be blocked. It is also better to use an alarm for trip signal monitoring
to block the reclosing operation in case the trip circuit continuity is detected by a logic circuit.
7) When a breaker fails to close. In case an auto-reclosing attempt is made but the breaker
stays open during the pre-calculated closing time, more reclosing attempts must be blocked.
8) Postpone the auto-reclosure to the after the arc extinction. This needs the application of
line-side voltage transformers [5].
9) For multi-phase faults, auto-reclosure methods must be blocked [5].
10) Perform reclosing operation immediately, if the trips are undesirable. Immediate re-
closing initiation can be performed by a substation SCADA computer if the dat from fault
shows there is a incorrect operation for relays.
11) If sync-check is employed for reclosing, change the sync-check angle. This angle can
be set if there is a considerable disturbance in the system. But, there might be not enough speed
ot capability of the algorithm to perform a fast angle calculation when there are rapid changes
in the system conditions [5].
Having used smart breakers, it is feasible to set the number of permitted reclosing attempts
or the reclosing time delay using the records of the operations of the breaker. As an example,
in cases there are too many operations in a limited time interval, it might be better to increase
the delay time length or to block the reclosing. But for faults with smaller currents might be
performed with shorter delays. For faults with high resistances it is also better to block the
reclosing. It is also possible for a computer to rotate the reclosure task for breaker-and-a-half
and ring bus structures.
Using adaptive reclosing, it will be possible to set reclosure mode for different lines. Using
WAN, wide area network, the dispatcher of the load can have communication with other sub-
stations and perform the online reclosure setting, i.e., using the current situation of the system.
Time-delayed auto-reclosing
For the time-delayed reclosing scheme application some points should be considered as
follows:
- Fault arc re-establishment probability reduction by increment in the predetermined time-

delay.
- Time-delayed reclosure is useful provided that adequate protection system is not applica-
ble for all of a transmission line terminals.
- The wanted re-establishment time of the system circuit, recognizing the number of termi-
nals of the line, stability studies, requirements of the system, kinds of the loads and damping
factors of the system.
Load-flow and stability studies must also be performed in case sync-check for unusually
long fault-clearance time is needed. Sometimes, double-checking of the phase difference and
being synchronous before application of the reclosure function is unnecessary due to the num-
ber of parallel paths with the faulted line.
Considering the time-delayed reclosure using the fastest reclosing schemes, usually only
one of the line terminals is being reclosed each time for re-energizing the line. Also, relays for
voltage monitoring are usually considered for determination of the system voltage presence on
one of the breaker sides. Therefore, the user will be able to select the conditions of the system
voltage in advance. Sometimes, users allow a delayed reclosing only when the phase difference
between the contacts of the open breaker in a predefined range and stays synchronous with for
a desired amount of time which is performed by sync-checks.
Stability considerations
While considering high-speed reclosure schemes to increase the transient stability margin
of the system, restore the important loads/customers, or restore some required interconnections
in the system, it must be noted that there are both merits and risks together. For example, if the
fault is permanent, there is absolutely no merit in performing the reclosing. For this purpose,
it is recommended to perform stability analysis on the system.
Switching surges
High-speed auto-reclosure schemes must not be employed if transient analysis studies on

the voltage show that there is a chance of switching surge production with magnitudes reaching
the design levels of the equipment.
Time considerations
Generally speaking, regardless of series capacitor applications, the initial reclosing dead-
time for the three-phase reclosure methods is between 0.5 to ten seconds for different systems
and conditions. In power system transmission lines, the time delay for reclosing usually cal-
culated by transmission line design engineers. For this purpose, stability analysis must be
considered so that the possible oscillations of the system after disturbances be damped, prop-
erly. Also, transient stability considerations of the system must be covered by auto-reclosing
function.
Single-shot and multiple-shot auto-reclosing
Single-shot reclosure means either high-speed or time-delayed reclosing the breaker during
a reclosing period, only for once. the other choice for reclosing, multiple-shot, is to reclose the
circuit breaker during the considered duty cycle for more than one time.
A typical reclosure at a line terminal may include a high-speed reclosing which is not su-
pervised by either sync-check or dead-line relaying, a time-delay reclosing which is supervised
by dead line voltage protection. A successful reclosing would allow the other line terminals
close by a time-delay, supervised by sync-check protection.
Generally speaking, modern high voltage and extra high voltage breakers are able to operate
in any multiple-shot reclosure period, practically. Some special issues regarding the circuit
breaker elements including opening or closing resistances, pressure of the available gas, air or
fluid, or factor of derating based on the expected fault current and number of fault interruptions
in a predefined time period in total can limit the effect on duty cycles or repetitive performances.
Before applying the multiple-shot reclosing, the user must make sure there would be no
danger of system instability or other possible limitations in case the fault is permanent. As
an example, multiple-shot reclosing is not practical when large motors or turbo-generators are
danger of mechanical damage is probable. Also, it is more desirable to postpone the second
reclosing shortly so that all transients of the previous reclosure are damped, properly.
Historically, single-shot reclosure of transmission lines has been in practice. But, later,
multiple-shot reclosure also became desirable and practical. This is mostly due to the fact
that most of the power system substations were not under supervisory control, traditionally.
In such substations, it is not practical by the personnel to perform the second reclosure after
one unsuccessful one. But, at substations where dispatching is performed by the personnel, an
unsuccessful single-shot reclosing may lead to a later outage which is not necessary.
However, at those locations where an employee must be dispatched, a delayed outage may
unnecessarily result if a single-shot reclosure is unsuccessful. About 25% and 10% of utilities
have records of successful 2nd and 3rd reclosing functions, respectively. But at the same time,
breaker constraints and reset time of the reclosing relays has to be taken into account not to
have problems in the result of additional operations of the breakers in a predefined time du-
ration. If there is no desire for single-shot reclosure schemes to be replace by multiple-shot
methods, a considerable success may be achieved by single-shot reclosing schemes just by in-
creasing the time-delay.
Single-phase tripping and auto-reclosing
The difference between single-phase and three-phase reclosing is that in single-phase re-
closing, tripping and reclosing is performed only to the faulted phase. In single-phase reclos-
ing, electrical power can be transfered through the non-faulted phases while the faulted phase
is open. This leads to decrement is rotor-angle drift rate between the synchronous machines
and increases the stability margin of the system. The only major problem is affection of the
ground protection of the system due to circulation of the ground currents.
During fault clearance in a single-phase reclosing, due to capacitive an also inductive cou-
pling of the phases, there is an induced voltage in the de-energized line. The magnitude of this
voltage depends on the mentioned capacitance values and therefore, on the transmission line’s
configuration and physical dimensions. This voltage may causes the arc to stay on for a longer
time period while the faulted line is isolated. In this case, the arc is called the secondary arc.
The main purpose of considering the reclosing time-delay is to give enough time to the sec-
ondary arc to be extinguished and therefore, to avoid reclosing-onto-fault and arc restriking.
The amount of this time-delay, called the dead-time, is chosen based on the system analysis.
One of the differences between single-phase and three-phase reclosing schemes is the differ-
ence in the dead-time which is considerably shorter in three-phase reclosing.
In some cases of reclosing, the amount of the induced voltage on the secondary arc is really
huge which can make the arc extinction time very long or even infinity. In such cases, one
practical method for decrement of the arc extinction time is to compensate the line capacitances
and therefore, to reduce the amount of the induced voltage. This is basically done using for-
legged shunt reactor, i.e., a three-phase reactor with a neutral reactor. Another possible method
is to close and reopen the grounding switches installed on the faulted phase in a fast manner.
In some special power system substation designing methodologies, single-phase tripping/reclosing
and three-phase tripping/reclosing are considered independent and non-related. In such de-
signs, it is the role of the protection system ro recognize the fault type, i.e., single-phase or
multi-phase and also to introduce the faulted phase(s). In general, three-phase tripping (with
no reclosing) are applicable when:
- There is a multi-phase fault,
- AG fault followed by an unsuccessful reclosure,
- Abnormal delay in performance of single-phase reclosure task while the fault is recog-
nized as AG,
- More phases become involved in the fault while the main faulted phase is isolated.
Blocking of auto-reclosing
There are different methods for blocking auto-reclosure schemes in transmission lines and
depend mostly on the design criteria of the power system. As examples of conditions when
reclosing should or must be blocked the following two items can be pointed:
- When there is a three-phase fault: Three-phase faults are not common on extra high
voltage systems (345 kV and above). But when there is a three-phase faults, it is not expected
to be temporary. Such faults are mostly caused by ground straps missed after the breaker
maintenance or downed line structures. This means it is better to block a three-phase reclosure
operation if there is a three-phase fault. As temporary three-phase faults are more likely for
transmission lines of lower voltage levels, auto-reclosure task might be interesting if generation
system stability are not affected, negatively.
- Out-of-step and power swing conditions: Auto-reclosing should be blocked when out-of-
step or power swing condition is detected. This is because auto-reclosing can further stimulate
an system which is already disturbed.
Multiple-breaker line terminations
Regarding multi-terminal transmission lines, assuming that for all terminals, the stability
of the system is not disturbed by reclosing on permanent fault and high-speed reclosing is
considered for all terminals, It is possible to apply high-speed reclosing. Generally speaking,
it is easier to apply high-speed reclosing to a two-terminal line than a multi-terminal one as at
least a shorter dead-time will be required.
When applying reclosing to a multi-terminal line, the pilot relaying channel has to be con-
sidered. Also, relays must be set for each of the terminals to watch for all possible faults at that
spot. This leads to more accurate performance as all terminals are being taken care all together.
However, it could be more interesting to perform high-speed reclosing at one of the terminals
and monitor other terminals for permanent fault detection.
Sometimes, it is practical to connect the line ends to more than one circuit breakers when
dealing with extra high voltage systems. Breaker-and-a-half, double-breaker and ring-bus con-
figurations are some examples for this purpose. For the fault clearance in such circuit arrange-
ments, more than a breaker needs to be tripped.
However, applying the reclosure task to all breakers at the same time can lead to more fault
duties. This means, the preference is to perform the line testing by applying reclosure to just
one of the breakers and then move to the rest in case of success. This can be double-checked
using sync-check and also by monitoring of the voltage. Sometime, the first reclosing is con-
sidered as high-speed.
Sync-check
Sync-check or synchronism-check is employed for supervision of electrical connecting two

parts of a system which are connected through ties in parallel with the path being closed [5].
Sync-check determines if the voltages of two side of the circuit breaker are closed enough,
and the frequency difference (slip) and the phase angle difference between them is within a
acceptable limit for a specific time interval. If the condition is not desirable, closing the circuit
breaker might have a harmful effect on equipment such as circuit breakers and shafts of turbo-
generators, and might also affect stability of the system as well. The setting of a sync-check
relay is performed to prevent reclosure when reclosing under live-line/live-bus conditions will
harmfully affect system operation and equipment.
Sync-check protection devices react to three main features related to the voltage phasors on
either side of the open breaker including magnitude difference, and phase angle difference and
slip. In some protection devices, specially microprocessor-based technologies, settings limita-
tion can independently be applied to each of the characteristics. In other protection devices,
specially old electromechanical designs, the characteristics are dependent so that separate lim-
itations can not be applied. characteristics.
Studies show that the effect of the reclosure is lower than closure into a fault for the most
of limits of sync-check angle employed by utilities and therefore, is not of concern. But, when
taking into account the possible effects on turbo-generator systems, it is desired to assess the
effect related to deliberate reclosure in a different manner than the effect of unavoidable faults.
The critical trade-off is to balance the possible effect related to the reclosure under violated
system conditions versus the possible effect on reliable operation of the system if reclosing is
prevented by the sync-check setting.
In case the angle difference of two sides of the breaker is not in the sync-check angle limit,
initially, it can be possible that the angle limit will be satisfied and met in a short time as a
results of changes to the system structure, e.g., reclosure in other transmission lines after a
considerable system disturbance or changes in generation dispatch. Logical circuit might dis-
able the auto-reclosure task in case appropriate auto-reclosure conditions are not met during a
predetermined time interval. The other possibility is that the logic might wait for acceptable
conditions to be met, indefinitely. Both methods have benefits and drawbacks. Limiting the
time interval in which a live-line/live-bus reclosing attempt is permitted will avoid having the
reclosing attempt performs in future in a potentially undesirable and unpredictable way. But,
this operation extends restoration of the system. The operation of indefinitely waiting will be
beneficial for decrement of the restoration time after a major disturbance by permitting reclo-
sure as soon as possible which is when the acceptable conditions exist. But, the unexpected
breaker closing during system restoration may disturb the generation and load balance. This
can lead to extend restoration time and subsequent outages. When this operation is employed,
it has to be based on analysis of the potential conditions of the system for which auto-reclosing
may happen for the purpose of managing the risk to system equipment and reliability. Regard-
less of which operation is employed, operators of the system are required to know how their
system is designed and need to be trained adequately to act during a system-wide outage event
in an appropriate manner.
Leader-follower auto-reclosing of transmission lines

The leader-follower auto-reclosure method is common practice in auto-reclosing in trans-
mission lines. Definition of the leader is the terminal of the line which performs auto-reclosing,
first, and the follower is the terminal of the line which performs the reclosing, second. This is
a different concept from the concept of lead-follow terminology employed for description of
different breakers of a multi-breaker terminal.
In leader-follower scheme, the weaker source of the two ends of a transmission line is
usually chosen as the leader end and is employed for testing of the line. This practice is
employed to verify the fault clearance when imposing the minimum amount of disturbances to
generation. But, it must be considered that the system testing from the weaker end might lead
to a larger drop of voltage and also disturb critical loads. If auto-reclose attempt at the leader
end is successful, auto-reclose attempt of the follower end which is supervised by sync-check
and/or voltage functions, is then enabled and auto-reclosing is performed. This technique may
be employed for restoration of either a tie line between two systems or a network line section.
A network line, i.e., a line at which the terminals are properly tied together by other parallel
lines, might employ just the voltage presence for initiation of auto-reclosure task while there are
adequate other network connections so that the difference of the voltage phase angles across the
open breaker is small. A tie line, i.e., a line where there are no other parallel lines connecting
the system tightly together, usually needs a sync-check relay to be applied so that the voltage
phase-angle difference across the breaker lies in appropriate ranges.
Auto-reclosing tasks in EHV systems are basically the same regardless of application of
series capacitors. But, overvoltage protection might be considered as well in case the voltage of
the follower terminal is too high to allow a safe auto-reclosure operation. In such conditions, a
transfer trip of the line’s leader terminal will be activated by the overvoltage protection system.
Chapter 3
Modeling and Behavior Analysis During

Single-phase Reclosing
In this Chapter, at first, modeling process of the studied power system is explained and then,
behavior of the modeled system during single-phase reclosing is analyzed. In the following, a
power system including a transmission line is modeled using PSCAD and the appropriate re-
sults are loaded to Matlab for more analysis. These results are used for performance evaluation
of the proposed reclosing methods.
The main aim of the performed power system analysis in this Chapter is to propose methods
to discriminate permanent and temporary faults based on the behavior of the faulted phase
voltage after single-phase opening of the circuit breakers. For this purpose, at first, behavior of
a simple three-phase transposed transmission line without shunt reactor is analyzed and then,
the analysis is extended to more general cases. In this order, some simplifications have been
made. As an example, resistances and inductances of equipment as well as source impedances

of both ends are ignored as they are negligible compared to line coupling impedances among
three phases [2].
54
3.1. Modeling of the Studied System 55
3.1 Modeling of the Studied System
3.1.1 Power system modeling
In this study, PSCAD is employed for modeling a typical 500kV power system in which con-
figuration of towers is set as horizontal. Single-line diagram of the network under study is
shown in Figure 3.1. In this study, voltage sources are considered as 1.02 p.u. (|V s | = 510kV)
and ∠V s = 21 Degrees for Side A and 1.0 p.u. (|V s | = 500kV) and ∠V s = 0 Degrees for Side B.
Length of the line is also 320 km while source impedances are chosen as 0.175 p.u. and 0.12
p.u. for the right and left voltage sources, respectively.
For the purpose of modeling of the transmission line, the Frequency Dependent (Phase)
model developed by Jose Marti is used [75]. In PSCAD two transmission line frequency depen-
dent models are offered, namely the Frequency Dependent (Mode) model and the Frequency
Dependent (Phase) model. It is recommended by PSCAD Users’ Guide that for all new studies
involving cables or transmission lines, the Frequency Dependent (Phase) model be employed.
This model was specifically added to PSCAD to replace the Frequency Dependent (Mode)
model. The Frequency Dependent (Phase) model can represent any type of transmission sys-
tem including aerial/underground and symmetrical/asymmetrical, and is more stable and more
accurate than Mode model [76].
The reason for choosing a 500kV transmission line for modeling and simulation is that it
is the highest voltage level in Canada at which single-phase reclosing is applicable [66]-[70].
The parameters of the system are also chosen as typical values for such transmission lines.
However, the developed methods in Chapters 4 and 5 are not dependent on system parameters,
including voltage level, because of the normalization and filtering of the information consid-
ered in the structures of the methods. The performances of the developed methods are also
verified using a real case study, a temporary fault case is a 765 kV transmission lne where the
solidity of the methods is shown.
56 Chapter 3. Modeling and Behavior Analysis During Single-phase Reclosing
In order to perform a better behavior analysis, a power system with three different topolo-
gies of transmission lines, including ideally-transposed, untransposed and partially-transposed
is used for simulation study and various permanent or transient faults inserted at proper loca-
tions. For the cases of ideally-transposed and untransposed lines, the line is divided into three
parts as two 30% parts at the sides and one 40% part at the middle. Therefore, the fault oc-
currence at 0%, 30%, 70% and 100% distances from one side of the transmission line can be
simulated.
For partially-transposed transmission line cases, the position of each conductor and/or line
configuration is changed during the study. In other words, the line is divided into three parts
with configurations of ’a-b-c’, ’c-a-b’, ’b-c-a’, which consist 50%, 30% and 20% of the line
length, respectively. With this selection, the entire line is not transposed 100% and therefore,
partially-transposed line is modeled. In the partially-transposed cases, the fault can occur at
both end and middle of each line section. Therefore, fault occurrence at 0%, 25%, 50%, 65%,
80%, 90% and 100% of the line from Side A is considered for simulation. It should be noted
that in the ’a-b-c’ horizontal configuration of towers, ’a’, ’b’ and ’c’ conductors are the left,
the middle and the right conductors, respectively. Importance of the partially-transposed case
is because in practice, it is not always possible to perform 100% transposition for transmission
lines, mostly due to technical limitations.
For the purpose of performance evaluation of the proposed reclosing methods for shunt
reactor compensated cases, a shunt reactor at each end with the three-phase reactive power
of 158.7 MVar are employed. Considering X/R ratio of 50, each reactor has been modeled
as Z = (31.5 + j1575.3) Ω. Neutral reactor is also represented by series combination of a

single-phase reactance and a resistance.
For the purpose of minimizing the secondary arc current, one common practice is to choose
a proper value for the neutral reactance, called optimum neutral reactor. Using a method ex-
plained in references [2, 27, 28, 83, 84] for this purpose, the optimum neutral reactor value is
VaF
Anti-aliasing
CVT VbF
Filter
VcF
R-L
Source
Src Imp Source
FD Line Src Imp
Impedance
Impedance
R-L
3ph 3ph
Side A 1ph Arc Side B

Model
TArc
VaF
Data Breaker
VbF
Recorder status
VcF
Figure 3.1: Single-line diagram of the modeled power system.
calculated as Xn = 260 Ω. Employing two similar shunt reactors at both line ends, a 520 Ω
(= 2 × 260 Ω) reactor is used for their neutrals. In cases the optimum neutral reactor is not
available, a larger/smaller reactor will be employed. In this study, neutral reactor impedances
equal to ZnLarge = (14 + j700)Ω, ZnOptimum = (10.2 + j510.8)Ω and ZnS mall = (6 + j300)Ω are used
to cover possible larger, optimum and smaller reactances in practice, respectively. Similar to
the shunt reactors, X/R ratios of all neutral reactors are assumed as 50 as well.
Also, for the cases of permanent faults, typical fault resistance values of 1 Ω, 10 Ω and 100
Ω are selected for various faulted cases. For performance evaluation of the proposed method in
presence of CVT transients, Lucas capacitive voltage transformer (CVT) model is used [85].
The modeled system in PSCAD with more details can be observed in Figures 3.2 and 3.3.
As observed in Figure 3.2, the line is divided into three sections namely TLine1, TLine2 and
TLine3 consisting 30%, 40% and 30% of the line, respectively. There are two four-legged
shunt reactors connected to buses BusS and BusR for the purpose of shunt compensation. For
uncompensated line cases, these shunt reactors must be ignored.
There is also a single-phase-to-ground fault applied to a bus employing a single-phase

breaker, called BreakerF. Therefore, modeling of fault occurrence at 30% distance from Side
A is shown in Figure 3.2. By applying similar faults at other buses of the system, it is possible
to simulate fault occurrence at different locations of the transmission line. In the performed
simulations, a constant value for the fault resistance, RF , is used for the permanent fault cases.
For the cases of temporary fault, the value of RF , which now is a current-controlled resistor, is
determined by the arc model to be explained in Subsection 3.1.2. At the moment of fault incep-
tion, the fault breaker closes and the arc resistor which has a very small value at that moment
becomes energized which leads to a huge fault current. However, as the time passes, the arc
resistance increases and therefore, the fault current decreases until becomes very small (less
than 2 A according to the standard [86]). At this moment, the arc is considered as extinguished.
The circuit breakers used at the left and right hand side of the circuit have the capability of
both three-phase and single-phase operation. Therefore, it is possible to perform appropriate
operations for permanent and temporary fault cases, including single-phase opening, single-
phase reclosing and three-phase opening. The blocks shown in the top-left side of the Figure
3.2 are models of CT and CVT used for the measurement purposes. At the bottom of CT and
CVT blocks, anti-aliasing filtering blocks are observed. The voltage and current data provided
during simulations are recorded using Comtrade 91 coding system [87].
There is a similar scenario about the circuit shown in Figure 3.3 in which the partially-
transposed transmission line is modeled. The only difference is combination of portions of the
line which are 50%, 30% and 20%. Therefore, the fault occurrence at 0%, 25%, 50%, 65%, 80
%, 90% and 100% distance of the line from Side A can be considered.
Figure 3.2: Single-line diagram of a power system with ideally-transposed and untrans-
posed transmission lines simulated in PSCAD.
Figure 3.3: Single-line diagram of a power system with partially-transposed transmission

line simulated in PSCAD.
3.1.2 Arc modeling
For the purpose of studying the reclosing in transmission lines, it is necessary to consider the
arcing nature of the temporary faults. Based on their features, electrical arcs can be categorized
into two groups, namely primary and secondary. A primary arc happens as the line insulator
string flashes over which is caused by reasons such as lightning strike. However, a secondary
arc, which is a more complicated phenomenon, appears following the primary arc while the
faulted phase is isolated due to breaker opening. For more information about electric arcs in
power systems see references [83]-[96].
There are few methods proposed for modeling of the secondary arc. In 1931, as the very
first attempt, Warrington introduced his empirical formula for arc modeling [97]. For this
purpose, he applied different voltage values across two electrodes with different distances and
provided some test results. In his work, having assumed a resistive nature for electrical arc,
he considered the relation between arc voltage and current as (3.1) where Ua and E a are arc
voltage and arc voltage gradient, respectively. L is the length of the arc while I is the arc current.
Parameters K and n are defined using the test results. The assumption of resistive nature for arc
was based on the fact that the arc voltages and the arc currents were in phase for each and every
of the test results. In [98]-[108] Terzija modified Warrington’s empirical formula considering
some assumptions, e.g., Warrington’s measurement devices were not accurate enough and the
arc length changes during the arc life. Terzija enhances the formula by considering the effect
of Gaussian noise.
Ua = E a L = (K/I n )L (3.1)
The most accurate method for modeling secondary arc is proposed by Kizilcay. He presents
a thermal model for secondary arc as (3.2) where garc is the conductance of the arc, τ is the
arc time constant, G is the stationary arc conductance, iarc is the arc current, u0 is the charac-
teristic voltage of the arc, r is the characteristic resistance of the arc per arc length, larc is the
instantaneous length of arc and t is the time variable [89]-[92].
dgarc /dt = 1/τ × (G − garc )

|iarc |
G= (u0 +r|iarc |)|larc |
(3.2)
Parameters u0 and r can be obtained based on measurements. For primary arcs, τ and larc
are constant and equal to their initial values τ0 and l0 , respectively, while they change by time
for secondary arcs. The time constant of the secondary arc is proportional to the arc length,
inversely, as defined in (3.3).
τ = τ0 × (larc /l0 )α (3.3)
In (3.3), τ0 and l0 are the initial time constant and length of the arc while α is a coefficient
in the range of 0.1 to 0.6. The length of the primary arc, which is the length of the flash-over,
is assumed to be 10% longer than the length of the line insulators string [92]. The length of
the secondary arc varies by time and is different for different wind speeds. For relatively slow
winds (slower than 1 m/s), the change of arc length is as (3.4) where l s (tr) is the time-dependent
length of the secondary arc with initial value of l0 and tr is the time duration starting from the
secondary arc initiation [86].
l s (tr )/l0 = 1, tr < 0.1
l s (tr )/l0 = 10tr , tr > 0.1 (3.4)

Table 3.1: Modeled Arc Parametrs.

Arc number τ0 (ms) l0 (m) u0 (v/cm)
1 0.714 3.5 12
2 0.555 3.15 8
3 0.833 3.5 11
In this study, secondary arc is modeled in PSCAD based on the model developed by Kizil-
cay using equations (3.2)-(3.4) [89]. For this purpose, the arc resistor shown as a block in
Figure 3.1 is considered as a current-controlled resistor. This means the value of the arc re-
sistor at each moment is considered to be in a non-linear relation with the arc current at that
moment. In should be noted that this relation is neither a function nor one-to-one as the history
of the arc and the initial conditions have the major roles in the arc behavior. In the performed
research study, three different sets of arc parameters available in Table 3.1 are employed for
providing needed arc characteristics using [96]. In Table 3.1, l0 and τ0 are the arc’s initial
length and time constant, while u0 is the characteristic voltage of the arc.
The detailed model of the arc used for simulation in PSCAD is shown in Figure 3.4. For a
better understanding of the arc model, it is easier to move in reverse direction, i.e., to start from
the output RF and reach the input iarc . In Figure 3.4, variable garc , the instantaneous value of
the arc conductivity is obtained by going back one step from output RF simply after a reversing
operation. Going one more step back leads to a variable equal to τ0 × dgarc /dt to be named
VRBL. This is because the passed element of the circuit is an integrator with coefficient value
of 1/τ0 . Using (3.2), VRBL equals to (G − garc ) × (τ0 /τ) and this can be confirmed in Figure 3.4
as VRBL is the result of division of two variables NU M and DNU M. However, NU M itself
equals to POS minus NEG which are equal to |iarc |/[(u0 + r|iarc |)|larc |] and garc , respectively,
using Figure 3.4. This means the value of NU M exactly equals G − garc according to (3.2). At
the other hand, DNU M equals to [(larc /l0 ])α which is exactly τ/τ0 using (3.3). Therefore, the
term VRBL = (G − garc ) × (τ0 /τ) is verified as VRBL = NU M/DNU M.
3.1.3 Protective relay modeling
A fourth order low-pass butter-worth anti-aliasing filter with cut off frequency of 1536 Hz is
used for filtering the outputs of the three-phase CVT models. The sampling rate for recording
the PSCAD results is also 20 kHz. The recorded data is imported to Matlab and decimated to
the sampling rate of 3840 Hz. Three phase voltages are passed through a CVT transient filter
proposed in [116] to obtain smoother waveforms for the phase voltages.
For the purpose of phasor estimation of the waveforms, including voltage and current wave-
forms, Cosine algorithm is used [130]. It must be noted that in theory, digital Fourier transform
filter, DFT, is the main filter to be used for phasor estimation. But, the output of this filter is a
combination of two different filters, called Sine and Cosine filters. However, it can be shown
that for systems with slow changes such as electric power system, the output of the Sine fil-
ter can be approximated by the output of the Cosine filter delayed by one quarter of a power
system cycle length. Therefore, it is a common practice in power system measurement and
relaying to use Cosine filter, instead of DFT. Protection pass of the system is also considered
as 1/16 of a cycle. Therefore, the protection algorithm is performed every 4 samples as each
power system cycle consists of 64 samples of data.
Figure 3.4: Details of the arc model simulated in PSCAD.

3.2 Behavior Analysis of the System During Single-Phase Re-
closing
3.2.1 Transposed line
Absence of shunt reactor
After single-phase opening of the circuit breaker in a transmission line, the secondary arc cur-
rent is fed due to capacitive and inductive coupling with other phases. Figure 3.5 (a) shows the
capacitive coupling circuit for a transposed line without shunt reactor. The simplified version
of the circuit is also shown in Figure 3.5 (b). In this figure, E x (x is h, k or s) is the average
of phase x voltages at both line ends (sending and receiving ends). It should be noted that this
simplification approach is common in secondary arc analysis as presented in [21] and is only
employed to derive the proposed adaptive reclosing methods while accurate simulations are
performed to verify the performances of the proposed techniques.
The inductive component of induced voltage to the faulted phase depends on load currents
in healthy phases as well as mutual inductive couplings between healthy and faulted phases
[22]. The inductive component could be modeled as an impedance in series with the equiv-
alent voltage source, i.e., E s , shown in Figure 3.5 (b). But, it is negligible and consistent
considering the fact that the healthy line load currents do not change considerably after line
isolation. Therefore, the inductive component is ignored in all the analyses in this Section.
3.2. Behavior Analysis of the System During Single-Phase Reclosing 67
Ek
Eh Sw
Es C0 (C1-C0)
Rarc
(a)
SW
+ (C1-C0)×2/3
Es VS Rarc C0 (Eh+Ek)/2
-
IS
(b)
Figure 3.5: (a) Equivalent circuit of a transposed line without shunt reactor in case of
single-phase reclosing of phase s, (b) rearranged circuit.
As per equivalent circuit shown in Figure 3.5, immediately after isolation of the faulted
phase, the phase voltage V s can be estimated as (3.5).
1
Rarc ||( jωC 0
) E + E
h k
Vs = 1 1
(3.5)
Rarc ||( jωC )+ 2
0 jω 23 (C1 −C0 )
where V s is the faulted phase voltage measured by the line protection relays; Rarc is the
arc resistance; C0 is zero sequence capacitance of transmission line; C1 is positive sequence
capacitance of transmission line; E h is the source voltage of the healthy phase (average of both
ends); E k is the source voltage of the healthy phase (average of both ends); ω is the angular
velocity; j is the complex operand and || stands for the impedance parallel operand.
Immediately after operation of the circuit breakers of both line ends, the arc resistance
is still small as compared to the line shunt capacitive impedance. Therefore, (3.5) can be
approximated to (3.6). According to (3.6), in case of low resistive faults, the magnitude of V s
is very small and the angle of V s leads (E h + E k ) by 90 degrees. This condition remains if there
is a permanent fault. However, for the case of a temporary fault, Rarc would increase, ideally
to infinity as the arc extinguishes. In this case, (3.5) could be approximated as (3.7). For a
typical transmission line, C1 is about 1.5 times of C0 [83]. As a result, |V s| would be about
12.5% of (E h + E k ) if arc is extinguished. In this case, V s would be in phase with (E h + E k ) if
arc resistance increases, considerably.
E
h +Ek
V s = jω 23 Rarc (C1 − C0 ) 2
(3.6)
1 1
if Rarc << jωC0
and Rarc << jω 32 (C1 −C0 )
2(C1 −C0 )
E
h +Ek 1
Vs = 2C1 +C0 2
if Rarc >> jωC0
(3.7)
To perform a better comparison, it is proper to define V sn , i.e., normalized V s , as per (3.8).

Magnitude and angle of V sn versus Rarc are shown in Figure 3.6 for the case of transmission
line without shunt reactor . As depicted, in case of a permanent fault where Rarc remains small
even after circuit beaker interruption, |V sn | will be close to zero while ∠V sn is almost 90◦ . In
case of a temporary fault where Rarc reaches a large value, |V s| could reach a value about 12.5%
depending on line positive and zero sequence capacitances. For such large values of arc resis-
tor, ∠V sn drops to zero. This can be confirmed using (3.7). In addition, for temporary faults, as
shown in Figure 3.6, both magnitude and angle of V sn change rapidly almost in a linear manner
versus arc resistance if Rarc is below a certain value (knee point in 3.6), e.g., about 900Ω in
this case. For arc resistance values above this number, both magnitude and angle of V sn change
considerably slowly until they settle down.

(a) (b)
100
∠ Vs (Degree)
|Vs| (pu)
0.1
50
0 0
0 2000 4000 6000 8000 0 2000 4000 6000 8000
Rarc Ohm Rarc Ohm
Figure 3.6: Normalized (a) magnitude and (b) phase angle of the faulted phase voltage.
V sn = V s /(E h + E k ) (3.8)
Regarding the above phasor analysis, it must be considered that this analysis is been used
only for obtaining the ideas and explanation of the developed concepts. In Chapters 4 and 5,
the methods are developed and evaluated not using phasor analysis, but transient simulation
results. However, there are two points regarding the validity of the phasor analysis for idea
development that must be indicated. The first point is that in the analysis of this Section, only
two states of the system are known for us, which are pre-fault and post-fault-clearance system
states. For obtaining the exact process in which the power system moves from pre-fault state
to post-fault-clearance state, a transient analysis must be performed. But, the results obtained
from the phasor analysis can be used as initial and final conditions. Also, considering the fact
that the main interest of the performed research study is power system protection, only the
fundamental frequency components of the voltage/current variables are of interest. Therefore,
results obtained from the phasor-analysis are valid for single-phase reclosing applications.
Presence of shunt reactor
In case of ideally-transposed transmission lines where shunt reactors are required, use of four-
legged shunt reactors is a common practice in industry to limit the secondary arc current for
faster and more successful single-phase reclosing scheme [20, 21]. The value of neutral reactor
can be selected as recommended in [2, 27, 28, 83, 84]. Equivalent circuit of a transposed
transmission line with four-legged shunt reactor is shown in Figure 3.7. As shown in Figure
3.7 (b), there is an inductive branch appeared in parallel with the capacitive branch. These two
inductive and capacitive branches would cancel out each other if the neutral reactor is selected
in such a way that parallel branches resonate. However, as the compensation is not 100% in
practice due to possible errors in line parameter estimation and other limitations discussed in
[62], eventually the equivalent circuit of the parallel branches would be either a large inductor
or a small capacitor. Therefore, share of the arc resistance from the feeding voltage, i.e.,
(E h + E k ), is even less than the case without shunt reactor, thus, the arc would be extinguished
faster.
Ek
Eh
Es Sw C0 L1
Rarc
(C1-C0) Ln
(a)
L12+3L1Ln
2Ln
(C1-C0)×2/3
SW
+
Es Vs Rarc C0 L1+3Ln
- (Eh+Ek)/2
Is
(b)
Figure 3.7: (a) Equivalent circuit of a transposed transmission line with four-legged shunt
reactor, (b) simplified circuit.
In this case, the measured voltage by the line protection relays can be estimated as (3.9).
Similar to the approach adopted earlier, immediately after operation of the circuit breakers of
both line ends when arc resistance is small, (3.9) can be estimated by (3.10). According to
(3.10), there is a linear relation between |V s| and Rarc in case of small arc resistances. Also,
the faulted phase voltage would either lead or lag the equivalent voltage source, (E h + E k ) by
90◦ depending on capacitive or inductive nature of X where X is the equivalent reactance of
the parallel branch. This would stay the case for permanent faults as fault resistance remains
small. However, for temporary faults, the arc resistance would increase to a considerably large
value. So, (3.9) could be estimated by (3.11) for such cases. Using (3.11), |V s| still is a factor of
(E h + E k ) although the factor is most probably smaller than the one in transposed lines without
shunt reactor. This is the case since X is typically a large value especially for cases at which
the neutral reactor is chosen properly. ∠V s also depends on capacitive or inductive nature of
nominator and denominator of (3.11). However, as both the nominator and the denominator are
imaginary reactances, (3.11) would have either a positive or a negative real value. Therefore,
V s would be either in phase or out of phase with (E h + E k ).
1
Rarc ||( jωC 0
)||( jω(L1 + 3Ln )) Eh + Ek
Vs = 1
( ) (3.9)
Rarc ||( jωC )||( jω(L1 + 3Ln ) + jX 2
0
It should be noted that the case of transmission line compensated by shunt reactors with
grounded neutrals is excluded in this research project. This is due to long arc extinction time
for this case. Therefore, single-phase reclosing is not recommended and is not practical in such
conditions [2].
Rarc Eh +Ek
Vs = jX
( 2 ) (3.10)
1
If Rarc << jωC0
and Rarc << jω(L1 + 3Ln )
1
( jωC )||( jω(L1 +3Ln )) E +E
Vs = 0
jX
( h2 k ) (3.11)
1
If Rarc >> jωC0
, Rarc >> jω(L1 + 3Ln )
1
and jωC0
|| jω(L1 + 3Ln ) << jX
3.2.2 Untransposed line
Absence of shunt reactor
The equivalent circuit of an untransposed transmission line without shunt reactor is shown in
Figure 3.8-(a). Simplified version of this circuit is also shown in Figure 3.8-(b). Based on the
circuit shown in Figure 3.8-(b), faulted phase voltage would be obtained as (3.12).
Rarc ||( jωC1 s n ) C sh E h + C sk E k

Vs = ( ) (3.12)
Rarc ||( jωC1 sn ) + 1
jω(C sh +C sk )
C sh + C sk
Comparing (3.5) and (3.12), V s in both cases depends on Rarc in similar ways and using (7),
it could be shown that voltage profiles similar to Figure 3.6-(a) would be obtained. The only
difference is that there would be a phase shift because of non-transposition. In other words,
E h and E k would have different participations to the equivalent voltage source for faults on the
side wire as C sh and C sk have different values.
Presence of shunt reactor
Equivalent circuits of the untransposed line and the simplified version of it in presence of four-
legged shunt reactor are shown in Figure 3.9. Using Figure 3.9-(c), the faulted phase voltage
would be as (3.13).
Rarc ||( jωC1 s n )||( jω(L1 + 3Ln )

Vs = E eq (3.13)
Rarc ||( jωC1 sn )||( jω(L1 + 3Ln )) + Xeq
where Xeq and E q are Thevenin equivalent reactance and source of the grayed circuit, re-
spectively.
Ek Chk
Eh Sw Csh Csk
Es Csn Chn Ckn

Rarc
(a)
SW (Csh+Csk)
+
Es Vs Rarc Csn (CshEh+CskEk)
-
(Csh+Csk)
IS
(b)
Figure 3.8: (a) Equivalent circuit of an untransposed transmission line without shunt
reactor, (b) simplified circuit.
Ek Chk
Eh Sw Csh C sk
Es Csn Chn Ckn L1

Rarc
Ln
(a)
L12+3L1Ln
2Ln
SW
+ (C sh+Csk)
Es (Eh+Ek)/2
VS Rarc L 1+3L n Csn (C shEh+CskEk)
-
(Csh+Csk)
IS
(b)
SW Xeq
+
Es VS Rarc L1+3Ln Csn
- Eeq
IS
(c)
Figure 3.9: (a) Equivalent circuit of an untransposed transmission line with four-legged
shunt reactor, (b) rearranged circuit, (c) simplified circuit.
Comparing (3.12) and (3.13) shows a great similarity in dependence of the faulted phase
voltage on the arc resistance. In other words, amplitude of the voltage increases in a similar
manner in both cases while arc resistance is increasing. Therefore, a similar voltage pattern,
close to Figure 3.6, is expected for this case as well. Again, similar to Subsection 3.2.1 for the
untransposed line without shunt reactors, there would be a different phase difference between
the voltages which is due to un-transposition.
3.3. Summary 75
3.3 Summary
In this Section, at first modeling process of the studied system, including the power system,
electric arc and the protection relay was explained. Then, it was shown that the general behav-
ior of the faulted phase voltage is somehow the same for all configurations of power systems.
In other words, when single-phase fault happens, voltage of the faulted point drops to rather
small values. This could remain the case if the fault is permanent. However, in case of tem-
porary faults, arc resistance starts increasing after the single-phase isolation of the line. This
would lead to increment of the faulted phase voltage to a considerable percentage of the rated
voltage once the arc is extinguished. This pattern of the faulted phase voltage magnitude and
angle behavior is the base for developing the proposed methods in this research study.
Chapter 4
Fault Type Recognition and Arc

Extinction Detection
In this Chapter, the voltage magnitude and phase angle patterns of the faulted phase voltage
after the line single-phase isolation are analyzed to perform adaptive single-phase reclosing.
For this purpose, two reclosing methods are proposed. In the first proposed adaptive single-
phase reclosing method to be called the angle-based method, the phase angle of the faulted
phase voltage is used for both fault type identification, i.e., permanent or temporary, and for
arc extinction detection in case the fault type is identified as temporary. In the second pro-
posed reclosing method to be called the phasor-based method, the first derivation of the faulted
phase voltage magnitude is used for the fault type recognition while the combination of first
derivation of angle and magnitude of the faulted phase voltage is used to quickly detect the
arc extinction for the temporary fault cases. Both proposed method are effective for shunt
compensated and uncompensated lines and for ideally-transposed, untransposed and partially-
transposed transmission lines.
76
4.1. Fundamentals of the Reclosing Methods 77
4.1 Fundamentals of the Reclosing Methods
In this Section, several simulated case studies are employed to verify the expected behavior
of the faulted phase voltage after the line single-phase isolation, derived in the Chapter 3. For
this purpose, permanent and temporary faults have been simulated and related waveforms are
presented. Voltage and current waveforms of the electric arcs/faults for six different temporary
fault cases are shown in Figures 4.1 to 4.6. For the all simulations, the single-phase-to-ground
fault inception time is 0.2s. The line single-phase isolation times are also 0.27s and 0.28s for
Sides A and B, respectively. Therefore, the faulted phase is completely isolated after t=0.28s.
(a)
1
|Varc|(pu)
−1
0.2 0.3 0.4(b) 0.5 0.6 0.7
10000
|Iarc|(A)
5000
0
−5000
0.2 0.3 0.4(c) 0.5 0.6 0.7
100
|Iarc|(A)
−100
0.3 0.4 0.5 0.6
Time (s)
Figure 4.1: Electric arc voltage and current for the case of a temporary fault in an ideally-
transposed transmission line, in absence of shunt reactor; (b) is zoomed in (c).
78 Chapter 4. Fault Type Recognition and Arc Extinction Detection
(a)
1
|Varc|(pu)
0
−1
0.2 0.3 0.4 (b) 0.5 0.6 0.7
15000
10000
|Iarc|(A) 5000
0
−5000
0.2 0.3 0.4 (c) 0.5 0.6 0.7
100
|Iarc|(A)
50
0
−50
0.3 0.4 0.5 0.6
Time (s)
Figure 4.2: Electric arc voltage and current for the case of a temporary fault in an ideally-
transposed transmission line, in presence of shunt reactor; (b) is zoomed in (c).
(a)
1
|Varc|(pu)
−1
0.2 0.3 0.4(b) 0.5 0.6 0.7
10000
|Iarc|(A)
5000
0
−5000
0.2 0.3 0.4(c) 0.5 0.6 0.7
100
|Iarc|(A)
−100
0.3 0.4 0.5 0.6
Time (s)
Figure 4.3: Electric arc voltage and current for the case of a temporary fault in an un-
transposed transmission line, in absence of shunt reactor; (b) is zoomed in (c).
(a)
1
|Varc|(pu)
0
−1
0.2 0.4 (b) 0.6 0.8 1
10000
|Iarc|(A)
5000
0
−5000
0.2 0.4 (c) 0.6 0.8 1
100
|Iarc|(A)
−100
0.3 0.4 0.5 0.6 0.7 0.8
Time (s)
Figure 4.4: Electric arc voltage and current for the case of a temporary fault in an un-
transposed transmission line, in presence of shunt reactor; (b) is zoomed in (c).
(a)
1
|Varc|(pu)
−1
0.2 0.3 0.4 0.5 0.6 0.7
(b)
15000
10000
|Iarc|(A)
5000
0
−5000
0.2 0.3 0.4 0.5 0.6 0.7
(c)
100
|Iarc|(A)
0
−100
−200
0.3 0.4 0.5 0.6
Time (s)
Figure 4.5: Electric arc voltage and current for the case of a temporary fault in a
partially-transposed transmission line, in absence of shunt reactor; (b) is zoomed in (c).
(a)
1
|Varc|(pu)
0
−1
0.2 0.3 0.4 0.5 0.6 0.7
(b)
15000
10000
|Iarc|(A) 5000
0
−5000
0.2 0.3 0.4 0.5 0.6 0.7
(c)
200
|Iarc|(A)
−200
0.3 0.4 0.5 0.6
Time (s)
Figure 4.6: Electric arc voltage and current for the case of a temporary fault in a
partially-transposed transmission line, in presence of shunt reactor; (b) is zoomed in (c).
4.1.1 Angle-based analysis
The faulted voltage phase angle, δ, the phase difference between the faulted phase voltage and
summation of the voltages of the non-faulted phases, as defined in (4.1), plays a major role in
the angle-based reclosing method. In (4.1), Vh = Vb , Vk = Vc and V s = Va in case of AG fault.
The negative signs in (4.1) are used to assign the pre-fault value of zero for δ and therefore, to
avoid δ values jumping from −180◦ to 180◦ as much as possible. As the phase angle measured
by Cosine algorithm is rotatory, a reference is needed in order to utilize the phase angle for
the proposed application. Considering the fact that the single-phase-to-ground fault is the only
interest of this research study, summation of voltage phasors of the non-faulted phases has been
chosen as a reference according to (4.1). This assumption also secures the proposed algorithm
against power system dynamics after line single-phase isolation since any system oscillation
appears almost equally in all three phase. Selection of this reference eliminates the effect of
angle variations due to system dynamics while it maintains the effect of arc extinction.
δ = ∠[−V s /(Vh + Vk )] (4.1)
Variable δmean , the mean value of the phase angle δ over the last power system cycle at each
time spot is the main variable employed in the angle-based method proposed in this research.
By averaging, the phase angle waveform becomes as smooth as possible which makes the
decision makings more reliable and more accurate.
Ideally-transposed line
Simulation results of two permanent fault cases in an ideally-transposed line are shown in
Figures 4.7 and 4.8. In both simulated cases, the fault resistance is equal to 1 Ω and the fault
occurs at 30% distance of the sending side, i.e., Side A. For the case shown in Figure 4.7, the
line is not compensated while shunt compensation is used for the case of Figure 4.8.
As observed, in the both cases δmean is around zero prior to the breakers’ single-phase
opening at t = 0.27 s. This means the faulted phase and summation of the two non-faulted
phase voltages (with negative sign) are still in phase. But after single-phase isolation of the
faulted line, δmean increases and after some transients finally settles down at a value around
90◦ to be called the-first-lock-angle, slightly after the single-phase isolation of the line. This
matches with what predicted in Section 3.2.
In Figures 4.9 and 4.10 two temporary fault simulated cases in an ideally-transposed line
without and with shunt compensation are shown, respectively. In both cases, the temporary
fault happens at 30% distance from the sending side. As observed, in both cases the variable
δmean is zero before the fault inception and slightly after the line isolation jumps to the-first-
lock-angle which is around 90◦ . This is similar to the previously explained permanent fault
cases of Figures 4.7 and 4.8. Variable δmean stays around the-first-lock-angle for a short period
of time which is because the arc resistor has a very small value right after the single-phase
isolation of the line. However, as the arc extinguishes and the arc resistor increases to very large
values, δmean increases to another value around 180◦ to be called the-second-lock-angle. This
means δmean turns almost out-of-phase respect to the equivalent voltage source of the faulted
phase after the arc extinction. In other word, an arcing/temporary fault is initially very similar
to a permanent fault when the arc resistor has a small value. However, as the arc resistance
increases, the difference between two fault types become more clear. This was also predicted
precisely in Section 3.2. The difference between the compensated and the non-compensated
cases is that there are some oscillations after the arc extinction at around t = 0.65 s for the
compensated case. These oscillations which happen between the shunt reactors and the line
capacitors do not exist for the un-compensated cases.
(a)
δmean (degree)
100
50
−50
0.2 0.3 0.4 0.5 0.6 0.7
(b)
95
(degree)
90
mean
δ
85
0.34 0.36 0.38 0.4
93
δmean (degree)
92
91
90
0.5 0.52 0.54 0.56 0.58 0.6
Time (s)
Figure 4.7: (a) δmean for the case of a permanent fault in an ideally-transposed transmis-
sion line, in absence of shunt reactor; (a) is zoomed in (b) and (c).
(a)
150
δmean (degree)
100
50
0
−50
0.2 0.3 0.4 0.5 0.6 0.7
(b)
110
(degree)
105
100
mean
δ
95
0.33 0.335 0.34 0.345 0.35
(c)
δmean (degree)
104
102
100
0.5 0.52 0.54 0.56 0.58 0.6
Time (s)
Figure 4.8: (a) δmean for the case of a permanent fault in an ideally-transposed transmis-
sion line, in presence of shunt reactor; (a) is zoomed in (b) and (c).
(a)
200
(degree)
100
mean
0
δ
0.2 0.3 0.4 0.5 0.6 0.7

(b)
120
(degree)
100
80
mean
δ
60
0.33 0.34 0.35 0.36 0.37 0.38
(c)
175
(degree)
170
mean
165
δ
0.54 0.56 0.58

Time (s)
Figure 4.9: (a) δmean for the case of a temporary fault in an ideally-transposed transmis-
(a)
200
δmean (degree)
100
0.2 0.3 0.4 0.5 0.6 0.7

(b)
105
δmean (degree)
100
95
0.33 0.335 0.34
(c)
(degree)
174
172
mean
δ
170
0.62 0.63 0.64 0.65
Time (s)
Figure 4.10: (a) δmean for the case of a temporary fault in an ideally-transposed transmis-
Comparing the simulation results, it is confirmed that the main difference between the per-
manent and temporary fault cases is in the final value of δmean variable after the arc extinction.
In fact, as predicted in Section 3.2, for both permanent and temporary fault cases, δmean variable
has values around 90◦ right after the line isolation. But, as the arc becomes extinguished for
temporary fault cases, δmean increases toward a number around 180◦ and settles down/resonates
around it while stays around 90◦ for the cases of permanent faults. This means, shortly after
the line isolation, δmean moves ideally 90 more degrees for temporary fault cases.
Untransposed line
δmean waveforms for two permanent fault cases in an untransposed line are shown in Figures
4.11 and 4.12. In both simulated cases, the fault resistance is equal to 1 Ω and happens at
30% distance of the sending side, i.e., Side A. For the case shown in Figure 4.11 the line is
not compensated while shunt compensation is used for the case of Figure 4.11. As observed,
in both cases δmean is around zero prior to the breaker opening at t = 0.27s. Similar to ideally-
transposed line cases, this means the faulted phase and summation of the two non-faulted phase
voltages (with negative sign) are still in phase. But after the single-phase isolation of the line,
δmean increases and after some transients, finally settles down around the-first-lock-angle equal
to 75◦ which is different from the-first-lock-angle associated with the ideally-transposed line
cases. This difference in the phase angle value after the single-phase isolation of the line is
because of the difference in phase angle of the equivalent voltage source behind the arc, which
itself is due to the asymmetry in the values of the coupling capacitors in untransposed lines.
(a)
100
(degree)
50
0
mean
δ
−50
0.2 0.4 0.6 0.8
(b)
80
(degree)
70
mean
δ
60
0.34 0.36 0.38 0.4
(c)
75
(degree)
74
mean
δ
73
0.5 0.55 0.6 0.65 0.7
Time (s)
Figure 4.11: (a) δmean for the case of a permanent fault in an untransposed transmission
line, in absence of shunt reactor; (a) is zoomed in (b) and (c).
(a)
100
δmean (degree)
50
−50
0.2 0.3 0.4 0.5 0.6 0.7
(b)
80
(degree)
75
mean
δ
70
0.34 0.36 0.38 0.4
(c)
δmean (degree)
74.4
74.2
0.5 0.52 0.54 0.56 0.58 0.6
Time (s)
Figure 4.12: (a) δmean for the case of a permanent fault in an untransposed transmission
line, in presence of shunt reactor; (a) is zoomed in (b) and (c).
(a)
δmean (degree)
150
100
50
0
−50
0.2 0.3 0.4 0.5 0.6 0.7
(b)
70
(degree)
65
mean
δ
60
0.33 0.34 0.35 0.36 0.37 0.38
(c)
138
(degree)
136
mean
134
δ
132
0.6 0.65 0.7
Time (s)
Figure 4.13: (a) δmean for the case of a temporary fault in an untransposed transmission
line, in absence of shunt reactor; (a) is zoomed in (b) and (c).
Similar to what discussed for the ideally-transposed cases, in untransposed line cases also,
the phase angle has similar behaviors right after the line isolation for permanent and temporary
faults. In fact, in untransposed lines, variable δmean is around 90◦ right after the line isolation.
But this variable starts increasing slightly after the line isolation as the arc is becoming extin-
guished. This can be seen in Figures 4.13 and 4.14 in which variable δmean is shown for tempo-
rary fault cases in absence and presence of shunt reactor, respectively. As seen in both figures,
the δmean value is zero before the line isolation, then jumps to the-first-lock-angle slightly after
the isolation and then, slowly moves toward the-second-lock-angle as the arc is extinguishing.
For uncompensated line cases, δmean settles down at the-second-lock-angle value after the arc
extinction while it resonates around it if the line is compensated.
(a)
150
(degree)
100
50
mean
0
δ
−50
0.2 0.4 0.6 0.8
(b)
δmean (degree)
60
50
40
30
0.34 0.36 0.38 0.4
(c)
123.5
δmean (degree)
123
122.5
0.8 0.85 0.9 0.95
Time (s)
Figure 4.14: (a) δmean for the case of a temporary fault in an untransposed transmission
line, in presence of shunt reactor; (a) is zoomed in (b) and (c).
Partially-transposed line
General behavior of δmean in partially-transposed lines is similar to ideally-transposed and un-

transposed line cases. In other words, this variable jumps from zero to the-first-lock-angle
slightly after the line isolation for the both fault types, stays there for permanent faults while
moves about 90 more degrees toward the-second-lock-angle for temporary fault cases. It is
almost predictable as partially-transposed line is the midway of ideally-transposed and un-
transposed lines. δmean waveforms for permanent and temporary fault cases in absence and
presence of shunt reactor are shown in Figures 4.15 to 4.18.
Considering the shown cases, for all three transmission line structures including ideally-
transposed, untransposed and partially-transposed, and for both line compensation conditions
including shunt compensated and uncompensated, they have a general pattern in common that
helps for the fault discrimination, i.e., discrimination between permanent and temporary fault
cases. In the all simulated cases, the phase angle δmean jumps from zero to the-first-lock-angle,
some values around 90◦ , shortly after the line isolation. This variable stays there for permanent
fault cases while it moves about 90 more degrees and reaches the-second-lock-angle for the
cases of temporary faults. This feature is used for fault type identification in the proposed
angle-based method.
If temporary fault is detected, the next step is to detect the arc extinction time as reclosing
on fault is so harmful for the power system stability and system equipment. In fact, the feature
of arc extinction detection is considered as very basic feature for an adaptive reclosing method.
For this purpose, behavior of the variable δmean must be monitored. In fact, as discussed earlier
in this Section, δmean either settles down at the-second-lock-angle, some values around 180◦ , for
un-compensated line case or resonates around it for the cases of compensated lines. Therefore,
the arc extinction signal will be activated when the settling down/oscillation is detected.
(a)
100
(degree)
50
mean
−50
δ
0.2 0.3 0.4 0.5 0.6 0.7
(b)
75
δmean (degree) 70
65
0.34 0.36 0.38 0.4
(c)
70.5
(degree)
70.45
mean
δ
70.4
0.5 0.52 0.54 0.56 0.58 0.6
Time (s)
Figure 4.15: (a) δmean for the case of a permanent fault in a partially-transposed transmis-
(a)
100
(degree)
50
0
mean
−50
δ
0.2 0.3 0.4 0.5 0.6 0.7

(b)
76
(degree)
75
74
mean
δ
73
0.34 0.36 0.38 0.4
(c)
74.7
(degree)
74.6
mean
δ
74.5
0.5 0.55 0.6 0.65 0.7
Time (s)
Figure 4.16: (a) δmean for the case of a permanent fault in a partially-transposed transmis-
(a)
200
(degree)
100
mean
0
δ
0.2 0.4 0.6 0.8
(b)
85
δmean (degree)
80
75
70
0.33 0.34 0.35 0.36
(c)
160
(degree)
158
mean
156
δ
0.54 0.56 0.58 0.6

Time (s)
Figure 4.17: (a) δmean for the case of a temporary fault in a partially-transposed transmis-
(a)
200
(degree)
100
mean
0
δ
0.2 0.4 0.6 0.8

(b)
90
(degree)
80
mean
δ
70
0.34 0.36 0.38 0.4 0.42 0.44
(c)
155
(degree)
150
mean
δ
145
0.6 0.65 0.7 0.75 0.8
Time (s)
Figure 4.18: (a) δmean for the case of a temporary fault in a partially-transposed transmis-
4.1.2 Phasor-based analysis
In the phasor-based proposed single-phase reclosing method, similar to Chapter 3, normalized

version of V s , i.e., V sn as defined in (3.8) is considered to make easier comparisons. In this
method, the major roles are being played by V snD , the first derivative of |V sn | as define in (4.2),
and by δD , the first derivative of δ. The derivation process for both variables are performed
using Least Error Square (LES) method. For this purpose, a window of data (here |V sn | or δ)
of length of a half a power system cycle is approximated by a straight line while the window
is moving toward positive numbers’ side in time axis. The amount of the derivative (V snD or
δD ) at each step is the slope of the straight line at that step. This process is done to guarantee
smooth waveforms for V snD and δD as any rapid change may lead to maloperation of the relay.
V snD = d(|V sn |)/dt (4.2)
Permanent fault
|V sn | and V snD for three different permanent fault cases are shown in Figures 4.19 to 4.21.
Waveforms shown in Figure 4.19 stand for a permanent fault case with a fault resistance of
1 Ω at 30% of the line from Side A in an ideally-transposed transmission line with no shunt
compensation. Figure 4.20 also represents the results for a permanent fault case while the
line is shunt-compensated untransposed and the fault resistance is 10 Ω. Figure 4.21 also
shows a permanent fault with 1 Ω resistance at 25% distance on an uncompensated line which
is partially-transposed. As shown for all three cases, |V sn | drops during fault, i.e., the time
interval between 0.2s and 0.28s. The amount of voltage drop depends on the fault location,
fault characteristics and system strength.
Once the line is single-phase isolated at 0.28s, |V sn | further drops to a considerably small
value and remains constant after transients due to system configuration change, filtering and
phasor estimation. This behavior of |V sn | exactly matches with the expected one derived by
the theoretical analysis presented in Section 3.2. After the line single-phase isolation, V snD
also drops smoothly from zero to its negative peak and then recovers again to zero as |V sn |
has already become steady. This negative peak of V snD , which will be used in the following
sections, is registered as V snDN . After line single-phase isolation, ideally V snD has no positive
peak as the faulted phase voltage is not recovered for permanent faults, while there can be a
small positive peak as observed in Figures 4.19-(c), 4.20-(c) and 4.21-(c), named V snDP , which
is because of system transients and filtering errors and V snD drops to zero after one power
system cycle in the worst case. It should be noted that presence of shunt reactor has no effect
on the waveforms in case of permanent faults because of the fact that shunt reactor draws
negligible current from the faulted phase due to the presence of small amount of faulted phase
voltage in both cases.
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 (b)0.4 0.5
20
0
VsnD
−20
−40
0.2 0.3 (c) 0.4 0.5
2
1
VsnD
0
−1
0.34 0.36 0.38 0.4
Time (s)
Figure 4.19: Normalized faulted phase voltage magnitude and its first derivative for a
permanent fault case in an ideally-transposed transmission line, in absence of shunt re-
actor; (b) is zoomed in (c).
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 0.4 0.5
(b)
20
0
VsnD
−20
−40
0.2 0.3 0.4 0.5
(c)
2
1
VsnD
0
−1
0.34 0.36 0.38 0.4
Tme (s)
permanent fault case in an untransposed transmission line, in presence of shunt reactor;
(b) is zoomed in (c).
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 (b)0.4 0.5
0
VsnD
−20
−40
0.2 0.3 (c)0.4 0.5

5
0
VsnD
−5
−10
−15
0.34 0.35 0.36 0.37
Tme (s)
permanent fault case in a partially-transposed transmission line, in absence of shunt re-
When the faulted phase voltage drops to zero due to the line single-phase isolation at t =
0.28s, V snD must ideally reach a considerable negative value and then recover to zero in a very
short time interval, as the voltage drops almost instantaneously. However, as observed in Figure
4.19, there is a time delay which is because of filtering, phasor estimation, and LES-based
derivation calculation delays each equal to 1.5, 1.25 and 0.5 cycles, respectively. Therefore, it
takes about 3.25 power system cycles for V snD to drop to V snDP and then recover to zero after
the line single-phase isolation.
Temporary fault
To validate the expected voltage behavior of the faulted phase derived in Section 3.2 for tem-
porary faults, several faults are simulated in ideally-transposed, untransposed and partially-
transposed lines, in absence and presence of shunt reactor, and at different fault locations. Also
for the purpose of arc modeling, Arc1 characteristics of Table 3.1 is used.
|V sn | and V snD for the six temporary fault cases are shown in Figures 4.22 to 4.27. As ob-
served, once the line is single-phase isolated at 0.28s, |V sn | reaches a considerably small value,
similar to the earlier permanent fault cases, and V snDN can be obtained similar to permanent
faults cases. However, voltage magnitude increases later as arc resistance increases. Compar-
ing Figures 4.22 to 4.27 with Figures 4.19 to 4.21, |V sn | of temporary fault cases recovers to
a higher value as compared to the permanent faults, after line single-phase isolation (t = 0.28
s). This pattern exactly matches with the expected behavior derived by the analytical analy-
sis presented in Section 3.2. V snD also makes a larger positive peak, to be called V snDP , after
voltage recovery compared to the permanent fault cases. In addition, V snD keeps holding large
positive values for a considerable time interval while the voltage is being recovered. This can
be observed by comparing Figures 4.22-(c), 4.23-(c), 4.24-(c), 4.25-(c), 4.26-(c) and 4.27-(c)
with Figures 4.19-(c), 4.20-(c) and 4.21-(c).
Regardless of the fault type, smaller |V snDN | and V snDP are obtained for the case of stronger
systems as the system resists against the faulted phase voltage change. This is also true for
the case of larger fault resistances as the fault is not that strong to change the system voltage.
However, the key point is that for the equal values of |V snDN |, larger V snDP values will be obtained
for temporary faults compared to permanent fault cases and in fact, there is a considerable gap
between V snDP values. This is actually true not only for V snDP , but also for all V snD values
after the second voltage recovery. This gap is even larger for V snD values after the positive
peak as V snD drops to zero considerably faster for permanent faults compared to temporary
fault cases. This is valid for all possible cases including ideally-transposed, untransposed and
partially-transposed transmission lines and even presence or absence of shunt and/or neutral
reactors has no effect on this fact. This feature is the main measure to discriminate temporary
and permanent faults in this research study.
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 0.4
(b) 0.5
20
0
VsnD
−20
−40
0.2 0.3 (c)0.4 0.5
3
2
VsnD
1
0
−1
0.35 0.4 0.45 0.5 0.55 0.6
Time (s)
temporary fault case in an ideally-transposed transmission line, in absence of shunt reac-
tor; (b) is zoomed in (c).
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 0.4 0.5
(b)
20
0
VsnD −20
−40
0.2 0.3 0.4 0.5
(c)
4
VsnD
0
0.35 0.4 0.45 0.5 0.55 0.6
Time (s)
temporary fault case in an ideally-transposed transmission line, in presence of shunt re-
(a)
1
|Vsn|(pu)
0.5
0
0.2 0.3 (b)0.4 0.5
0
VsnD
−20
−40
0.2 0.3 (c)0.4 0.5
2
VsnD
1
0
−1
0.35 0.4 0.45 0.5 0.55 0.6
Time (s)
temporary fault case in an untransposed transmission line, in absence of shunt reactor;
(a)
|Vsn|(pu)
0.5
0
0.2 0.4 0.6 0.8
(b)
20
VsnD
0
−20
−40
0.2 0.4 0.6 0.8
(c)
2
VsnD
1
0
−1
0.4 0.6 0.8
Time (s)
temporary fault case in an untransposed transmission line, in presence of shunt reactor;
(a)
|Vsn|(pu)
0.5
0
0.2 0.3 0.4 0.5
(b)
20
0
VsnD
−20
−40
0.2 0.3 0.4 0.5
(c)
2
VsnD
−2
0.35 0.4 0.45 0.5 0.55 0.6

Time (s)
temporary fault case in a partially-transposed transmission line, in absence of shunt re-
(a)
|Vsn|(pu)
1
0.5
0
0.2 0.3 0.4 0.5 0.6 0.7
(b)
20
0
VsnD
−20
−40
0.2 0.3 0.4 0.5 0.6 0.7
(c)
2
VsnD
−2
0.35 0.4 0.45 0.5 0.55 0.6 0.65

Time (s)
temporary fault case in a partially-transposed transmission line, in presence of shunt
reactor; (b) is zoomed in (c).
Arc extinction
The other vital point about temporary fault cases to consider is how to recognize arc extinction
as the breaker is not supposed to reclose before real arc extinction. This is done in the phasor-
based method based on the fact that after the arc extinction, some kind of stability appears in
the system. Therefore, both V snD and δD variables, i.e., the first derivations of |V sn | and δ, either
settle down at or resonate around their final values which is zero for both variables.
In the PSCAD simulations, the arc current is monitored to determine the actual arc extinc-
tion time. The time instance when the current is above 2A for the last time is recorded as the
actual arc extinction time [84]. Having done this for 275 temporary fault cases, it is observed
that the arc extinction time estimated using the above-mentioned features is very close to the
actual arc extinction time.
For the purpose of arc extinction detection, both variables of V snD and δD are employed. δD
for two temporary fault cases in an untransposed line is shown in Figures 4.28 and 4.29. In
these figures, δD settles down (Figure 4.28-(b)) or resonates (Figure 4.29-(b)) at t = 0.6s and
t = 0.82s, respectively, while the arc extinction times using V snD are t = 0.55s and t = 0.75s.
Comparing the obtained arc extinction times with PSCAD results of t = 0.57s and t = 0.78s, it
has been observed that δD settles down or resonates slightly later than reality while V snD reacts
faster than the actual arc. This was the case for almost all of the temporary fault simulated
cases. In the phasor-based reclosing method, both V snD and δD are used for the arc extinction
detection purpose.
(a)
10
5
δD
−5
−10
0.2 0.4 0.6 0.8
(b)
2
1
δD
−1
0.5 0.6 0.7 0.8 0.9
Time (s)
Figure 4.28: First angle derivative for the case of a temporary fault case in an untrans-
posed transmission line, in absence of shunt reactor; (a) is zoomed in (b).
(a)
10
δD
0
−5
−10
0.2 0.4 0.6 0.8
(b)
1
0.5
δD
0.75 0.8 0.85 0.9

Time (s)
Figure 4.29: First angle derivative for the case of a temporary fault case in an untrans-
posed transmission line, in presence of shunt reactor; (a) is zoomed in (b).
4.2 Proposed Adaptive Reclosing Algorithms
4.2.1 Angle-based algorithm
Block diagram of the proposed angle-based adaptive single-phase reclosing algorithm is shown
in Figure 4.30. The angle-based algorithm can be initiated either by auxiliary contacts of the
line circuit breakers or through analyzing CT secondary current as proposed in [19]. The
faulted phase is determined and the process starts if single-phase-to-ground fault inception is
confirmed. Present-day line protection relays are equipped with phase selector function which
reliably determines the faulted phases.
According to the shown diagram in Figure 4.30, voltage values of all three phases and the
breaker interruption information are used for calculation of the δ parameter after anti-aliasing
and CVT temporary filtering and phasor estimation [116]. Having δ in hand, δmean is obtained
using an averaging unit. This averaging unit is simply a combination of an accumulator and a
divider, together a digital filter.
4.2. Proposed Adaptive Reclosing Algorithms 101
The obtained δmean variable is used for fault type identification. In case the detected fault
type is permanent, three-phase trip signal will be issued. But in case temporary fault is identi-
fied, single-phase reclosing signal will be issued after arc extinction detection. Arc extinction
process itself is performed using δmean variable too.
va vb vc
Breaker Pre-filtering and

Interuption Phasor Estimation
Calculator
Averaging
mean
Permanent Fault Type Temporary

Recognition
Arc Extinction
Detection
Final decision making

(Tripping/Reclosing)
Figure 4.30: Block diagram of the proposed angle-based adaptive single-phase reclosing
algorithm.
Fault type recognition
As mentioned in Subsection 4.1.1, similarity between permanent and temporary faults is that
for both cases, δmean moves from zero to an angle around 90◦ , called the-first-lock-angle. The-
first-lock-angle is dependent on the line structure and the faulted phase, but in an ideally-
transposed line, it is 90◦ , ideally. The difference between the two fault types projected to
δmean is that for permanent faults, δmean stays at the-first-lock-angle while it moves about
more 90 degrees for temporary fault cases, either in positive or negative direction, and set-
tles down/resonates around another phase angle value called the second-lock-angle. Therefore,
it is possible to recognize the fault type in two simple steps as:
1) To find the-first-lock-angle,
2) To realize if δmean stays around the-first-lock-angle or moves toward another value, i.e.,
the second-locked-angle.
In the proposed algorithm, δmean is monitored after the line isolation to see if it stays around
any value with variation as small as 5◦ for one full length of the power system cycle. This
angle value is detected as the-first-lock-angle. In the next step, δmean is monitored for 5 more
power system cycles. The fault type will be detected permanent if δmean stays in the range, i.e.,
at the-first-lock-angle ±2.5◦ , for at least five power system cycles. Otherwise, it is a temporary
fault.
Arc extinction detection
For temporary fault cases, the next step after the fault type identification is arc extinction de-
tection so that reclosing onto fault is prevented. The detected time for the arc extinction must
be slightly after the real arc extinction, as too long arc extinction detected time means the
system will have to work in unbalanced condition, unnecessarily. For this purpose, two differ-
ent cases are possible, shunt compensated and uncompensated lines. For the both cases, the
variable δmean increases toward the-second-lock-angle. It will stay there in case the line in not
compensated while it will resonate around the-second-lock-angle for shunt compensated lines.
In the proposed arc extinction detection algorithm, technically, there are two blocks that
can detect the arc extinction, settling-down detection and resonance detection. The initial duty
of the settling-down detector is to detect the second-locked-angle. This happens when δmean
increases to a value and stays around ±2.5◦ for at least 3 full cycles. This can be done by
simply monitoring the δmean variable. The other possible case is when δmean resonates around
the second-locked-angle. In such a case, at first the peak and the next dip right afterward
around the second-locked-angle are detected. Then, the very first time at which δmean reaches
the average value of the peak and the dip is recognized as the arc extinction time.
4.2.2 Phasor-based algorithm
Block diagram of the proposed phasor-based reclosing algorithm is shown in Figure 4.31. Sim-
ilar to the angle-based reclosing method, the phasor-based algorithm can also be initiated either
by auxiliary contacts of the line circuit breakers or through analyzing CT secondary current as
proposed in [19]. The faulted phase is determined and the process starts if single-phase-to-
ground fault inception is confirmed. Present day line protection relays are equipped with phase
selector function which reliably determines the faulted phases. In the proposed algorithm, as-
suming that the faulted phase is known, V snD and δD are obtained by derivating |V sn | and δ by
means of half-cycle LES algorithm, after anti-aliasing and CVT temporary filtering and pha-
sor estimation [116]. Using these variables, first, the fault type is identified. For the permanent
fault cases, three-phase tripping will be initiated immediately, while single-phase reclosing will
be commanded once arc extinction is detected, for temporary faults.
v a vb vc
|Vsn | δ
Anti-aliasing Filter LES Derivation LES Derivation
+
CVT Transient Filter Breaker VsnD δD
+ interruption
Phasor Estimation detection Fault type recognition
(for transient faults)
Determine Vh, Vk, Vs Final decision making
Figure 4.31: Block diagram of the proposed adaptive single-phase reclosing algorithm.
Fault type recognition
As mentioned in Subsection 4.1.2, after the line single-phase isolation, the ratio of V snD to
|V snDN | is larger for temporary faults than permanent fault cases and the difference becomes
even larger after the positive peak, V snDP . Hence, the voltage derivative ratio, V DR, as given in
(4.3) is defined for fault identification.
V snD
V DR = | | (%) (4.3)
V snDN
To calculate V DR, V snD must be monitored after the line single-phase isolation to obtain
V snDN . Considering the fact that it takes around 3.25 cycles for V snD to make a negative peak
and then recover to zero after the line single-phase isolation, as mentioned in Subsection 4.1.2,
V snDN must be obtained within this time interval. For permanent fault cases, V snDP appears
slightly after 3.25 cycles and then V snD drops to zero shortly, while it will remain positive as
the arc is being extinguished in cases of temporary fault. Therefore, by monitoring V snD for a
sufficient time interval after the line single-phase isolation, the fault type can be identified.
In this research study, considering various delays such as local circuit breaker opening
detection, delay between circuit breaker operation at both line ends, CVT temporary and relay
filtering and estimations, it is proposed to monitor V snD for 6 cycles or 100ms after the line
single-phase isolation. Then, V DR is calculated at each protection pass for the 6th cycle, i.e.
the time interval within 83ms to 100ms. In this case, 16 values for V DR will be obtained for the
6th cycle. The fault type will be recognized using V DRt , the summation of the 16 obtained V DR
values. In cases of permanent fault, V DRt is ideally zero although in practice it can have a very
small value due to various sources of error. In cases of temporary fault, V DRt is considerably
larger than the one obtained for permanent faults. In this research study, V DRt = 5% has been
chosen as a threshold for discrimination. For temporary fault detection, V DR must be larger
than 5%, while, V DR < 5% leads to permanent fault detection.
In the case of a close-in fault with very small resistance, voltage drops to zero immediately
after the fault inception. Therefore, line single-phase isolation does not lead to further voltage
drop. This means V snDN will be detected as zero and this will be the case for all 16 values of
V snD of the 6th monitored cycle. To cover this condition, permanent fault is detected in case the
faulted phase voltage magnitude stays less than 0.5% of the rated voltage for the entire 100ms
monitoring time interval.
In case of temporary faults, the main sign of arc extinction is some kind of settlement in system
variables. In this case, both V snD and δD settle down at/resonate around their final values as
the arc extinguishes and they are chosen for arc extinction detection purpose as discussed in
Subsection 4.1.2.
Based on the comprehensive simulation study performed, it was observed that in most of
the cases, the signs of the arc extinction appear in V snD about or before the real arc extinction,
whereas, these signs show up in δD always after the real arc extinction. Further, in case of fast
arc extinction, it is hard to find a proper reference time to start monitoring δD for arc extinction
detection as there are large transients on δD after the line single-phase isolation as shown in
Figures 4.28 and 4.29. In addition, the arc extinction algorithm requires to analyze V snD or δD
for several samples after real arc extinction to reliably detect the arc extinction. In this research
study, it is proposed to use both quantities to increase the reliability and speed of detection.
In the proposed method, having recognized the temporary fault, initial arc extinction is de-
tected if V snD becomes less that 10% of V snDP for three consecutive protection passes. Then,
this initial arc extinction is confirmed by observing δD smaller than a very small value; 0.1◦ /s
is considered in this research study. This strategy increases the security of the reclosing algo-
rithm and also the reclosing speed by eliminating the waiting time for arc extinction detection.
The merit of involving V snDP in the arc extinction detection algorithm is to provide an adaptive
threshold for V snD settlement detection as faster extinguishing arcs result in larger V snDP . There-
fore, it is not needed to wait for V snD settlement detection for too long in such cases, while for
slower arcs, the algorithm will wait longer to ensure reliable detection as V snDP is smaller.
4.3 Simulation Results
In this research, 550 simulated cases have been used for performance evaluation of the pro-
posed method. The simulated cases can be categorized into three groups based on the trans-
position method considered for the transmission line including ideally-transposed (100 cases),
untransposed (210 cases) and partially-transposed (240 cases). Various cases are different in
terms of fault location, arc type (Arcs 1 to 3 for temporary fault cases as per Table 3.1), arc
resistance (for permanent fault cases) and presence/absence of shunt/neutral reactor.
4.3.1 Angle-based method
According to the algorithm of the angle-based reclosing method, variable δmean is used for fault
type recognition. In this order, if this variable stays at the-first-lock-angle ±2.5◦ for 5 power
system cycles, then the fault is considered as permanent, otherwise it is a temporary fault and
therefore, the next step would be to find the arc extinction time to activate the reclosing signal.
4.3. Simulation Results 107
Average fault type detection time, i.e., temporary or permanent, for transmission lines with
ideally-transposed, untransposed and partially-transposed structures are available in Tables 4.1
to 4.3. In these tables, fault location is the distance of the fault inception point from the sending
end, in percent of the line length. Using these tables, the average time for ideally-transposed
and untransposed lines is from 9ms to 233ms, and from 7ms to 221ms, respectively. This is 7
ms to 194ms for fault type detection in the partially-transposed line.
Average needed time for detection of a temporary fault in ideally-transposed, untransposed
and partially-transposed lines are 69ms, 88ms and 81ms, respectively. The same values with
the same order for permanent fault detection are 155 ms, 122 ms and 150 ms. Average time
needed for detection of a temporary fault including all temporary fault cases is 91ms while it
is 139ms for all permanent fault cases. As observed, there is considerably longer time needed
for permanent fault type detection than for temporary. The reason is that the algorithm has to
wait for 5 power system cycles to make sure if it is a permanent fault, while it can detect a
temporary fault any time after the-first-lock-angle as soon as the sign of the temporary fault
appears, i.e., increment of the variable δmean more than 2.5◦ .
Considering the compensation condition of the lines, regardless of the fault type, the aver-
age needed time for the fault detection is 106ms and 115ms for the ideally-transposed line in
absence and presence of shunt reactor, respectively. For the untransposed line, it is 95ms and
108ms and for the partially-transposed line, it is 83ms and 92ms, respectively. This means, the
proposed algorithm has a better performance for uncompensated lines.
Table 4.1: Average temporary fault detection time for an ideally-transposed and an un-
transposed line using the angle-based method (ms).
Fault location 0% 30% 70% 100%
Ideally No shunt reactor 9 76 76 72

Transposed With shunt reactor 43 101 77 69
Side conductor No shunt reactor 73 89 86 79

Untrans- With shunt reactor 68 78 48 66
posed Middle conductor No shunt reactor 7 75 74 74
With shunt reactor 68 78 48 66
Table 4.2: Average permanent fault detection time for an ideally-transposed and an un-
transposed lines using the angle-based method (ms).


Table 4.3: Average fault type recognition time for a partially transposed transmission line
using the angle-based method (ms).
Fault location 0% 25% 50% 65% 80% 90% 100%
Temp- No shunt reactor 7 73 71 70 71 68 70

orary With shunt reactor 102 83 79 79 85 82 84
Perm- No shunt reactor 194 146 144 142 140 140 143
anent Wth shunt reactor 193 147 145 144 143 141 142
The average fault detection time including all the simulated cases for ideally-transposed,
untransposed and partially-transposed lines is 224 ms, 105 ms and 92 ms, respectively. This
means the algorithm has the fastest operation for the partially-transposed line while it is the
slowest for the ideally-transposed line. This can be a good sign as partially-transposition is the
most practical structure of the lines while there is no ideally-transposed line in practice. And
finally, the average fault detection time including all the simulated cases is 140ms.
Average fault type recognition times, including both temporary and permanent, for ideally-
transposed, untransposed and partially-transposed lines in absence and presence of shunt re-
actor are shown in Figures 4.32 to 4.35. As observed, for faults not so close to the relay
location, the temporary fault recognition speed is considerably close for ideally-transposed
and untransposed line cases. This is also true for permanent fault type cases. However, for
partially-transposed line, the fault recognition speed is considerably faster, for both tempo-
rary and permanent fault cases. Also, in general, the performance of the method is faster for
uncompensated lines than for compensated lines considering all three line configurations.
IA IP UA UP
Fault detection time (ms) 100
50
0
0 30 70 100
Fault location (%)
Figure 4.32: Average temporary fault detection time for an ideally-transposed and an
untransposed transmission line in Absence (A) and Presence (P) of shunt reactor using
the angle-based method (ms).
IA IP UA UP
Fault detection time (ms)
200
100
0
0 30 70 100
Fault location (%)
Figure 4.33: Average permanent fault detection time for an ideally-transposed and an
untransposed transmission line in Absence (A) and Presence (P) of shunt reactor using
the angle-based method (ms).
A P
Fault detection time (ms) 100
50
0
0 25 50 65 80 90 100
Fault location (%)
Figure 4.34: Average temporary fault detection time for a partially-transposed transmis-
sion line in Absence (A) and Presence (P) of shunt reactor using the angle-based method
(ms).
A P
Fault detection time (ms)
200
150
100
50
0
0 25 50 65 80 90 100
Fault location (%)
Figure 4.35: Average permanent fault detection time for a partially-transposed transmis-
sion line in Absence (A) and Presence (P) of shunt reactor using the angle-based method
(ms).
After the fault type recognition, a proper decision must be made by the relay. For permanent
fault cases, the proper decision is always to issue the three-phase trip signal as there is no hope
for the fault to be cleared by itself. However, there is a chance for fault clearance if the fault
is temporary. For such cases, the next step is to issue the single-phase reclosing signal. But,
reclosing has to happen in a reasonable amount of time after the arc extinction. If reclosing
happens too late, the system has worked in an unbalanced condition for some amounts of time,
unnecessarily, and if it happens before the arc extinction, it is harmful for the system equipment
and the system stability. For this purpose, an adaptive method must be able to detect the exact
arc extinction time.

Average and maximum reclosing time delay, the time distance between the arc extinction
detection moment proposed by the adaptive angle-based algorithm and the real arc extinc-
tion time obtained based on PSCAD simulations, for ideally-transposed, untransposed and
partially-transposed line simulated cases are presented in tables 4.4 to 4.6. For each fault loca-
tion and/or transmission line configuration, the averaging is applied to three time delay values
obtained for three sets of arc parameters, including faster/medium/slower arcs.
In table 4.4, the average reclosing time delay for the ideally-transposed case is between 48
and 59ms. The same values for untransposed and partially-transposed cases are 45 to 114ms
and 18 to 74ms, respectively. Therefore, the overall performance of the method is between 18
and 114ms including all temporary fault cases. However, reclosing delay is between 70 and
160ms for the method proposed in [62]. Method of [56] also provides the average process-
ing times between 29 and 67ms for temporary faults. Therefore, the method of this research
study has been faster than the method of [62] but slower in total than the method proposed in
[56]. It might be considered that the proposed method in [56] is only applicable for ideally-
transposed lines. In addition, it employs data of both sides of the transmission line and requires
communication facilities too, while the proposed technique in this research study is based on
local data application. The proposed technique is also effective for any kind of system trans-
position configuration and absence/presence of shunt reactor has no effect on the technique’s
performance.
The average arc extinction detection delay of the proposed phase angle-based algorithm
for the ideally-transposed line is 49 ms and 57 ms in absence and presence of shunt reactor,
respectively. For the untransposed line, these values are 60ms and 74ms. Also, they are 46ms
and 50ms for the cases of the partially-transposed line. This shows that the proposed algorithm
can detect the arc extinction faster when the line is not compensated.
Regardless of the shunt compensation condition of the line, the arc extinction detection
delay of the algorithm is 55ms, 70ms and 49ms respectively for the ideally-transposed, un-
transposed and partially-transposed lines. This means the proposed algorithm shows the high-
est performance for the partially-transposed line while the slowest performance belongs to the
untransposed line. Finally, the overall delay of the algorithm in detection of the arc extinction
is 58ms.
Table 4.4: Average reclosing time delay for an ideally-transposed and an untransposed
lines using the angle-based method (ms).


Table 4.5: Maximum reclosing time delay for an ideally-transposed and an untransposed
lines using the angle-based method (ms).


Table 4.6: Average and maximum reclosing time delay for a partially transposed trans-
mission line using the angle-based method (ms).
Fault location 0% 25% 50% 65% 80% 90% 100%
Ave- No shunt reactor 54 62 60 48 35 21 41

rage With shunt reactor 74 89 71 53 36 18 18
Max- No shunt reactor 61 70 67 58 47 36 52

imum Wth shunt reactor 89 104 78 61 63 32 45
Average and maximum reclosing time delay for ideally-transposed, untransposed and partially-
transposed lines are shown in Figures 4.36 to 4.39. First of all, it is observed that average and
maximum reclosing time delay values are considerably close to each other for each fault lo-
cation/line configuration. This means, for each fault location/line configuration, the proposed
method is able to detect the arc extinction after almost the same time delay regardless of the
arc extinction speed/arc parameters. Hence, it can be concluded that the proposed method is
not considerably sensitive to the arc parameters.
It is also observed that, beside the close-in fault cases, the fault location does not have
specific effect on the average and maximum reclosing time delays for ideally-transposed and
untransposed lines. Also, the method has had a better performance for ideally-transposed line.
However, for partially-transposed line cases, average performance of the method increases by
increment of the fault distance from the relay location. This means, for the faults at further
locations, the method has better performances. However, the maximum reclosing delay is
smaller for faults at the middle of the line.
IA IP UA UP
Reclosing time delay (ms)
80
60
40
20
0
0 30 70 100
Fault location (%)
Figure 4.36: Average reclosing time delay for an Ideally-transposed (I) and an Untrans-
posed (U) transmission lines in Absence (A) and Presence (P) of shunt reactor using the
angle-based method (ms).
IA IP UA UP
Reclosing time delay (ms) 150
100
50
0
0 30 70 100
Fault location (%)
Figure 4.37: Maximum reclosing time delay for an Ideally-transposed (I) and an Untrans-
angle-based method (ms).
A P
80
60
40
20
0
0 25 50 65 80 90 100
Fault location (%)
Figure 4.38: Average reclosing time delay for a partially-transposed transmission line in
Absence (A) and Presence (P) of shunt reactor using the angle-based method (ms).
A P

80
60
40
20
0
25 50 65 80 90 100
Fault location (%)
Figure 4.39: Maximum reclosing time delay for a partially-transposed transmission line
in Absence (A) and Presence (P) of shunt reactor using the angle-based method (ms).
The proposed algorithm has also been applied to a field recorded data, a temporary AG
fault in 765 kV Rockport substation in American Electric Power grid [62]. The faulted line is
untransposed compensated by shunt reactor with switching technology only at one end [21, 22].
Time domain waveforms of the normalized faulted phase voltage magnitude, |V sn |, and
the mean value of its phase angle, δmean are shown in Figure 4.40. The fault inception time
and the line single-phase isolation time are 100ms and 177ms, respectively. Using the event
recorder time reference, reclosing occurs at 630 ms. As per the depicted waveforms, the fault
is a temporary fault which becomes clear at about 300ms.
The proposed algorithm recognizes the-first-lock-angle as 99.3◦ at t = 219ms. Then, the

temporary fault is detected at t = 249 ms where δmean does not stay around the-first-lock-
angle and moves toward negative values. Finally, the resonance of δmean around −50 Degrees
is detected at t = 326ms which means the arc is extinguished. The developed arc extinction
technique is 304ms faster than the actual reclosing and 84ms faster than the technique proposed
in [62]. Therefore, the proposed algorthm has a successful performance for the considered field
recorded temporary fault case.
(a)
|Vsn|(pu)
0.5
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(b)
(Degree)
100
0
δmean
−100
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(c)
(Degree)
100
δmean
90
0.2 0.21 0.22 0.23 0.24 0.25
(d)
(Degree)
0
−50
δmean
−100
0.3 0.32 0.34 0.36 0.38 0.4
Time (s)
Figure 4.40: Waveforms of a temporary fault case in Rockport 765 kV station in Amer-
ican Electric Power grid, (a) normalized voltage magnitude of the faulted phase and (b)
average value of the faulted phase voltage angle; (b) is zoomed in (c) and (d).
4.3.2 Phasor-based method
According to Subsection 4.2.2, the fault type is recognized by means of V DRt values. For
all simulated cases, the gap between V DRt values for temporary and permanent faults is large
enough. In fact, the minimum V DRt value for temporary faults is 20%, while it is never larger
than 0.5% for any permanent fault case. The average V DRt values for temporary and permanent
fault cases are also 180% and 0.15%, respectively. The proposed algorithm has had successful
performance for all simulated cases. The fault type for all cases is recognized 100ms after the
line single-phase isolation. This is considerably faster than the method introduced in [62] with
permanent fault detection times within 220ms to 280ms and almost equal to the communication
aided technique proposed in [56] with detection times within 99ms to 108ms.
The average and maximum reclosing time delay, the time distance between the proposed
algorithm arc extinction detection moment and the real arc extinction time obtained based on
PSCAD simulations, for ideally-transposed, untransposed and partially-transposed line simu-
lated cases are presented in tables 4.7 and 4.8. For each fault location and/or transmission line
configuration, the averaging is applied to three time delay values obtained for three sets of arc
parameters, including faster/medium/slower arcs. It is observed that the average and maximum
time delays and/or individual time delays related to the each arc parameter set are pretty close
for each fault location/line configuration. This means, for each fault location/line configura-
tion, the proposed method is able to detect the arc extinction after almost the same time delay
regardless of the arc extinction speed. Hence, it can be concluded that the proposed method is
not considerably sensitive to the arc parameters.
Table 4.7: Average/maximum reclosing time delay for an ideally-transposed and an un-
transposed lines using the phasor-based method (ms); Sh.R.: Shunt Reactor; S.C./M.C.:
Fault on Side/Middle Conductor.
Ideally No Sh.R. 89 / 93 91 / 103 91 / 103 81 / 85

Transposed With Sh.R. 39 / 44 40 / 43 40 / 43 32 / 39
S.C. No Sh.R. 90 / 94 83 / 97 100 / 111 97 / 105

Untrans- With Sh.R. 86 / 91 76 / 86 62 / 92 65 / 86
posed M.C. No Sh.R. 92 / 97 91 / 99 75 / 90 83 / 100
With Sh.R. 72 / 91 70 / 86 68 / 92 59 / 86
Table 4.8: Average and maximum reclosing time delay for a partially transposed trans-
mission line using the phasor-based method (ms); Sh.R.: Shunt Reactor.
Fault location 0% 25% 50% 65% 80% 90% 100%
Ave- No Sh.R. 76 85 77 61 51 45 56
rage With Sh.R. 49 60 42 32 31 42 60
Max- No Sh.R. 78 88 87 63 60 58 67
imum Wth Sh.R. 60 75 50 33 33 49 70
In tables 4.7 and 4.8, the average reclosing time delay for temporary fault cases is 32 to 91
ms while this between 70 and 160ms for the method proposed in [62]. Method of [56] also
provides the average processing times between 29 and 67ms for temporary faults. Therefore,
the method of this research study has been faster than the method of [62] but slightly slower in
total than the method proposed in [56]. It might be considered that the proposed method in [56]
is only applicable for ideally-transposed lines. In addition, it employs data of both sides of the
transmission line and requires communication facilities too, while the proposed technique in
this research study is based on local data application. The proposed technique is also effective
for any kind of system transposition configuration and absence/presence of shunt reactor has
no effect on the technique’s performance.
Average and maximum reclosing time delay for ideally-transposed, untransposed and partially-
transposed lines are shown in Figures 4.41 to 4.44. Similar to the angle-based method, for the
phasor-based method is also observed that average and maximum reclosing time delay values
are considerably close to each other for each fault location/line configuration. This means,
for each fault location/line configuration, the proposed method is able to detect the arc extinc-
tion after almost the same time delay regardless of the arc extinction speed. Therefore, the
conclusion is that the proposed method is considerably insensitive to the arc parameters.
IA IP UA UP
100
Prediction time error (ms) 80
60
40
20
0
0 30 70 100
Fault location (%)
Figure 4.41: Average reclosing time delay for an Ideally-transposed (I) and an Untrans-
phasor-based method (ms).
IA IP UA UP
Prediction time error (ms)
100
50
0
0 30 70 100
Fault location (%)
Figure 4.42: Maximum reclosing time delay for an Ideally-transposed (I) and an Untrans-
phasor-based method (ms).
A P
80
60
40
20
0
0 25 50 65 80 90 100
Fault location (%)
Figure 4.43: Average reclosing time delay for a partially-transposed transmission line in
Absence (A) and Presence (P) of shunt reactor using the phasor-based method (ms).
A P
80
60
40
20
0
0 25 50 65 80 90 100
Fault location (%)
Figure 4.44: Maximum reclosing time delay for a partially-transposed transmission line
in Absence (A) and Presence (P) of shunt reactor using the phasor-based method (ms).
It is also observed that for the ideally-transposed and untransposed line cases, the fault
location has no considerable effect on the reclosing delay while for the partially-transposed
cases the method has a better performance when the fault happens at the middle of the line.
(a)
1
ag (pu)
0
|V |
−1
0.1 0.2 0.3 0.4 (b) 0.5 0.6 0.7 0.8
1
sn (pu)
|V |
0.5
0
0.1 0.2 0.3 0.4 (c) 0.5 0.6 0.7 0.8
10
0
VsnD
−10
−20
0.1 0.2 0.3 0.4 (d) 0.5 0.6 0.7 0.8
20
δD
0
−20
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Time (s)
Figure 4.45: Waveforms of a temporary fault case in Rockport 765 kV station in Amer-
ican Electric Power grid, (a) normalized instantaneous voltage waveform of the faulted
phase, (b) normalized voltage magnitude of the faulted phase and (b) its first derivative,
(c) first angle derivative.
Similar to the angle-based method, performance of the phasor-based algorithm is also eval-
uated using a field recorded data, a temporary AG fault in American Electric Power grid [62].
The fault happened in Rockport substation on a 765 kV untransposed line which employs
shunt reactor with switching technology only at one end [21, 22]. Instantaneous waveform,
magnitude and the first derivation of the faulted phase voltage as well as the first derivation
of its phase angle are shown in Figure 4.45(a) to 4.45(d). The fault inception time and the
line single-phase isolation time are 100ms and 177ms, respectively. Using the event recorder
time reference, reclosing also occurs at 630 ms. As per the depicted waveforms, the fault is a
temporary fault which becomes clear at about 300ms.
The proposed algorithm correctly recognizes the fault type based on the calculated V DRt =
150% at t = 280 ms or 100 ms after the line single-phase isolation. Then, according to the
algorithm of the phasor-based method, the arc extinction process has two steps. At first, there
is an initial arc extinction detection based on V snD and then, this has to be confirmed by δD .
Using the proposed method, the initial arc extinction has been detected at 300ms. This is the
time at which V snD becomes less than 10% of V snD ’s positive peak for the first time after line
single-phase isolation and stays there for three samples as shown in Figure 4.45 (c). Then, this
is confirmed at 320ms at which δD becomes negative as shown in Figure 4.45 (d). So, the final
arc extinction time obtained using the algorithm is 320ms. This is 310ms faster than the actual
reclosing and 90ms faster than the technique proposed in [62].
4.3.3 Comparison of the reclosing methods
Two proposed adaptive single-phase reclosing methods, i.e., the angle-based and the phasor-
based methods, have different structures, algorithms and logics and therefore, they will have
different performance in different situation. However, using all simulated cases, both methods
have had a successful operation and the difference is only in operation speed.
In term of structure of the methods, the angle-based method has a simpler structure as uses
only the phase angle of the faulted phase while the phasor-based method uses the faulted phase
voltage magnitude and derivatives of both magnitude and phase angle of the voltage.
In therms of fault type recognition, the logic of the angle-based method is based on imme-
diate/late detection of temporary/permanent faults. In fact, the default fault type is permanent
in this method and this is double checked throughout a predetermined time period of length of
five power system cycles or 83ms after the first-angle-locked detection, while temporary faults
can be detected anytime in that period of time. In fact, the average time needed for detection
of a temporary/permanent fault including all simulated cases is 91/139ms for the angle-based
method. Therefore, the angle-based method has a faster fault type detection performance for
temporary faults than permanent faults.
However, the phasor-based method has no default fault type and has has to wait until the
end of its 100ms time period for the fault type recognition. But, the time period is shorter than
the period used for the angle-based method. Therefore, for temporary faults, the angle-based
method has had a faster performance while phasor-based method acts faster for permanent fault
cases. But, the fault type recognition speed is more important for permanent fault cases than
for temporary faults. The reason is that after the permanent fault detection, the next step is
three-phase tripping while the recloser is to wait for the arc extinction for temporary faults.
Therefore, the phasor-based method has a better performance in fault type recognition.
In terms of reclosing time delay, the angle-based method is able to detect the arc extinction
18 to 114 ms after the arc extinction while it is 32 to 91 ms for the phasor-based method.
Therefore, the time delay values are more diverse for the angle-based method than the phasor-
based. Considering all temporary fault simulated cases, the angle-based method has a reclosing
time delay not longer than 59ms while the phasor-based method can do the same job not later
than 80ms after the real arc extinction, in average. Therefore, when the fault is temporary, the
phase-angel-based method acts faster at both fault detection and arc extinction detection.
As a comparison of the performances of the methods with the existing adaptive reclos-
ing methods, two of the best existing methods were chosen, including one local and one
communication-based method, using [62] and [56]. The local method was able to detect the
arc extinction 70 to 160ms after the line isolation while the communication-based method was
faster and could do the same job in 29 to 67ms. Therefore, the proposed methods, which are
local, were faster than the local method, while they were slightly slower the communication-
based method, in average. Therefore, the improvement is observed in their performances com-
pared to the existing reclosing methods.
4.4 Summary
In this Chapter, two methods for adaptive single-phase reclosing in transmission lines were
proposed. The first method, called the angle-based method, uses phase angle of the faulted
phase, both to recognize the fault type, i.e., permanent or temporary, and to detect the arc
extinction in case the fault is temporary. The second method, the phasor-based method, uses
the first derivative of the faulted phase voltage magnitude for fault type recognition, and the
first derivatives of both magnitude and angle of the faulted phase voltage, for arc extinction
detection.
Both methods had successful performances for transmission lines with various transposi-
tion configurations, including ideally-transposed, untransposed and partially-transposed. Com-
pensation condition was not also an important factor to affect the performance of the methods.
Both methods were evaluated using 550 different simulated cases as well as a real case
study, and the results confirmed the capabilities of the methods. Both methods had a successful
performances for all of the simulated cases. The angle-based method had a better performance
for temporary fault cases while the phasor-based method acted better when the fault type was
permanent. In terms of the structures of the methods, the angle-based method has a simpler
structure and also, can be considered as the more reliable method as used the data itself and
not the derivative of the data. Performances of the proposed methods was also compared with
two of the best existing adaptive reclosing methods, one local and one communication-based.
Methods had better performances than the local method while they were slightly slower than
the existing communication-based method.
Chapter 5
Arc Extinction Time Prediction
When single-phase reclosing function is considered, what normally happens for a temporary
fault case is to open the faulted phase and reclose the circuit breaker after a delay. The tra-
ditional single-phase reclosing uses a predetermined time delay to avoid arc restriking. This
time delay must be chosen properly as reclosing the breaker when temporary fault has not been
cleared yet can further damage the equipment and put the system stability at risk.
Adaptive single-phase reclosing methods have been proposed in the literature to identify
the fault type including temporary and permanent, and then, to recognize if the arc is extin-
guished [55]. However, all the proposed adaptive single-phase reclosing methods have to wait
for the arc to be extinguished if the fault is temporary. The problem is that, in some cases, the
arc extinction speed is too low and it is harmful for the system to serve with one phase open
for a long time which leads to issues such as injection of negative sequence component to the
system. Therefore, it would be beneficial to predict the arc extinction time for temporary faults
well in advance to initiate three-phase tripping signal in case of a slow extinguishing arc. How-
ever, prediction of the arc extinction time after breaker interruption has not been investigated
in the literature.
127
128 Chapter 5. Arc Extinction Time Prediction
In this Chapter, voltage magnitude of the faulted phase after single-phase isolation of the
line is used to predict the arc extinction time for temporary faults. For this purpose, there are
two methods proposed and the results are compared. The proposed methods are effective for
both uncompensated and compensated transmission lines with shunt reactor and are consider-
ably insensitive to transposition condition of the line.
5.1 Fundamentals of The Proposed Prediction Methods
To verify the special behaviors of the faulted phase voltage during arc extinction, six simulated
temporary fault case studies are used in this Section. In all six simulated cases, a temporary
single-phase-to-ground fault incepts at t = 0.2s. At t = 0.27s and t = 0.28s, breakers of Sides

A and B isolate the faulted line by performing single-phase opening. Therefore, after t = 0.28s
the faulted phase is fully isolated. For performing an easier and more clear comparison, similar
to Chapter 4, V sn and V snD variables are employed in this Chapter as well, as defined in (3.8).
5.1.1 Analysis of the faulted phase voltage waveforms
Waveforms of arc voltage and current for six different temporary fault cases on an ideally-
transposed, an untransposed and a partially-transposed line in absence and presence of shunt
reactor are shown in Figures 5.1 to 5.3. As observed in all three figures, the arc voltage is very
small immediately after fault inception, i.e., t = 0.2s, while the arc current is very large. How-
ever, the arc voltage increases and the arc current drops to small values as the arc extinguishes
gradually, after single-phase opening of the breakers at t = 0.28s.
5.1. Fundamentals of The Proposed Prediction Methods 129
(a) (b)
1 1
arc (pu)
arc (pu)
|
|
0 0
|V
|V
−1 −1
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.3 0.4 0.5 0.6 0.7
(c) (d)
15000
10000 10000
arc (A)
arc (A)
5000 5000
|
|
0 0
|I
|I
−5000 −5000
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.3 0.4 0.5 0.6 0.7
(e) (f)
100 100
arc (A)
|Iarc|(A)
50
|
0
0
|I
−100 −50
0.3 0.4 0.5 0.6 0.3 0.4 0.5 0.6
Time (s) Time (s)
Figure 5.1: Electric arc voltage and current for the case of two temporary faults in an
ideally-transposed transmission line in {(a) and (c)} absence and {(b) and (d)} presence of
shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively.
(a) (b)
1 1
arc (pu)
arc (pu)
|
0 0
|V
|V
−1 −1
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.4 0.6 0.8 1
(c) (d)
10000 10000
arc (A)
arc (A)
5000 5000
|
0 0
|I
|I
−5000 −5000
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.4 0.6 0.8 1
(e) (f)
100 100
|Iarc|(A)
arc (A)
|
0 0
|I
−100 −100
0.3 0.4 0.5 0.6 0.3 0.4 0.5 0.6 0.7 0.8
Time (s) Time (s)
Figure 5.2: Electric arc voltage and current for the case of two temporary faults in an
untransposed transmission line in {(a) and (c)} absence and {(b) and (d)} presence of
shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively.
(a) (b)
1 1
arc (pu)
|Varc|(pu)
|
|V 0 0
−1 −1
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.3 0.4 0.5 0.6 0.7
(c) (d)
15000 15000
10000 10000
|Iarc|(A)
|Iarc|(A)
5000 5000
0 0
−5000 −5000
0.2 0.3 0.4 0.5 0.6 0.7 0.2 0.3 0.4 0.5 0.6 0.7
(e) (f)
200
100
|Iarc|(A)
arc (A)
0
|
0
−100
|I
−200 −200
0.3 0.4 0.5 0.6 0.3 0.4 0.5 0.6
Time (s) Time (s)
Figure 5.3: Electric arc voltage and current for the case of two temporary faults in a
partially-transposed transmission line in {(a) and (c)} absence and {(b) and (d)} presence
of shunt reactor; (c) and (d) are zoomed in (e) and (f), respectively.
Waveforms of |V sn |, magnitude of V sn , and V snD , associated with the six simulated temporary
fault cases are shown in Figures 5.4 to 5.6. As seen in Figures 5.4 to 5.6, (a) and (b), during the
time interval between the fault occurrence moment and single-phase opening of the beakers,
i.e., t = 0.2s to t = 0.28s, |V sn | decreases to some extent depending on system strength and also
characteristics and location of the fault. Following isolation of the faulted line at t = 0.28s,
|V sn | drops considerably and then, is recovered slowly by increment of the arc resistance. At the
same time, as observed in Figures 5.4 to 5.6, (c) and (d), V snD drops to a negative number after
single-phase line isolation, recovers to zero and registers a positive peak to be called V snDPeak as
shown in Figures 5.4 to 5.6, (e) and (f). V snD stays above the horizontal axis for a considerable
time interval during the voltage recovery and eventually resonates/settles down around zero in
presence/absence of shunt reactor after the arc extinction.
5.1. Fundamentals of The Proposed Prediction Methods 131
(a) (b)
1 1
|Vsn|(pu)
|Vsn|(pu)
0.5 0.5
0 0
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
(c) (d)
20 20
0 0
VsnD
VsnD
−20 −20
−40 −40
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5
(e) (f)
3 4
2 VsnDPeak
V
VsnD
snD
Peak
snD 2
1
V
0 0
−1
0.35 0.4 0.45 0.5 0.55 0.6 0.35 0.4 0.45 0.5 0.55 0.6
Time (s) Time (s)
Figure 5.4: Normalized faulted phase voltage magnitude and its first derivative for the
temporary fault cases in an ideally-transposed transmission line shown in Figure 5.1; (c)
and (d) are zoomed in (e) and (f), respectively.
(a) (b)
1 1
|Vsn|(pu)
|Vsn|(pu)
0.5 0.5
0 0
0.2 0.3 0.4 0.5 0.2 0.4 0.6 0.8
(c) (d)
20
0 0
VsnD
VsnD
−20 −20
−40 −40
0.2 0.3 0.4 0.5 0.2 0.4 0.6 0.8
(e) (f)
2 2
VsnDPeak VsnDPeak
VsnD
VsnD
1 1
0 0
−1 −1
0.35 0.4 0.45 0.5 0.55 0.6 0.4 0.6 0.8
Time (s) Time (s)
temporary fault cases in an untransposed transmission line shown in Figure 5.2; (c) and
(d) are zoomed in (e) and (f), respectively.
(a) (b)
1 1
sn (pu)
sn (pu)
|V |
|V |
0.5 0.5
0 0
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5 0.6 0.7
(c) (d)
20 20
0 0
VsnD
VsnD
−20 −20
−40 −40
0.2 0.3 0.4 0.5 0.2 0.3 0.4 0.5 0.6 0.7
(e) (f)
2 VsnDPeak 2 VsnDPeak
snD
snD
0 0
V
V
−2 −2
0.35 0.4 0.45 0.5 0.55 0.6 0.35 0.4 0.45 0.5 0.55 0.6 0.65
Time (s) Time (s)
temporary fault cases in a partially-transposed transmission line shown in Figure 5.3; (c)
and (d) are zoomed in (e) and (f), respectively.
For a temporary fault case in which single-phase reclosing of the line is considered, V snD
waveform can be approximated by a decaying exponential curve for the arc extinction time
interval, i.e., from the moment at which V snD = V snDPeak to the arc extinction moment. By this
choice, it is possible to predict the arc extinction time by finding the zero-crossing point of the
approximating exponential curve.
5.2 Proposed Prediction Methods
In this Section, two different methods are proposed for the purpose of prediction of the arc
extinction time during single-phase reclosing, namely, the point-based and the window-based
methods. In the following, the proposed methods are explained, the prediction results are
compared, and merits and drawbacks of the methods are expressed.
5.2. Proposed Prediction Methods 133
5.2.1 Point-based method
Exponential curve approximation for the point-based method
A decaying exponential function is defined as x = ke−t/T where k and T are magnitude and
decaying time constant, respectively, where T determines the decaying speed. Both k and T
can be obtained having the data of only two points of the curve, e.g. (t1 , x1 ) and (t2 , x2 ) using
(5.1) where x1 = ke−t1 /T and x2 = ke−t2 /T . Having k and T , to , the time spot associated with any
desired xo value can be calculated as (5.2).
x1 /x2 = e(t2 −t1 )/T => (5.1)

t2 −t1
T= Ln(x1 /x2 )
and k = x1 /e(−t1 /T )
to = Ln(k/xo )T (5.2)
Proposed algorithm for the point-based method
According to Figures 5.5, (e) and (f), V snD increases from negative to positive values as the
arc is being extinguished after the line isolation. Then, it makes a positive peak and decreases
toward zero. After the arc extinction, V snD either settles down at zero for the uncompensated
line cases, or resonates for the lines with shunt/neutral reactor. It is desired that the point-based
method be able to make decision within 100 ms after the single-phase line isolation. In this
order, if the after-peak part of the V snD curve is approximated by a decaying exponential curve,
it will be possible to predict the arc extinction time. Having the predicted arc extinction time,
one of the possible choices, reclosing or three-phase tripping will be selected.
For obtaining the approximating exponential curve, all needed is the information of two
points of the curve as discussed in the first paragraph of this Subsection. The first chosen
point is the point of the V snD peak, called V snDPeak . The second chosen point is the point of
the minimum value for V snD after the peak point and before the decision making time, called
V snDMin . It was mentioned in the previous paragraph that the decision is made within 100ms
after the single-phase line isolation of the faulted phase. Applying the information of these two
points to (5.1), parameters k and T of the approximating exponential curve and therefore, the
constants of the approximating exponential equation are available.
V snD curves drawn in Figure 5.5 associated with two temporary fault cases of Figure 5.2, as
well as the related approximating curves using the point-based prediction method, are shown
in Figure 5.7.
(a) (b)
2 VsnD VsnD
VsnDP VsnDP
VsnDMin 2 VsnDMin Approximating
Approximating
1 exponential exponential
1
0 0
−1 −1
0.3 0.4 0.5 0.6 0.7 0.8 0.3 0.4 0.5 0.6 0.7 0.8
Time (s) Time (s)
Figure 5.7: V snD and the approximating exponentials associated with the temporary fault
cases shown in Figure 5.5.
The last step is to obtain the estimated arc extinction time, the time at which the approx-
imated V snD becomes zero. But, an exponential curve becomes zero at t = ∞, theoretically.
Therefore, a threshold must be set for V snD , i.e. V snDT h , and the predicted arc extinction time
will be the time at which the approximated V snD becomes smaller than the threshold V snDT h . In
this research, two values for the threshold have been chosen, V snDT h = 0.1% and V snDT h = 1%.
At the end, the chosen parameter must be applied to (5.2), i.e. xo = V snDT h , to obtain the
predicted arc extinction time.
5.2. Proposed Prediction Methods 135
5.2.2 Window-based method
Exponential curve approximation for the window-based method
An exponential curve is defined as (5.3) where k is the magnitude, T is the decaying time
constant and c is the DC offset [131]. For negative values of T , the exponential curve eventually
approaches the dc offset c. Equation (5.3) can be simplified and rearranged to the form of a
linear equation, i.e., y=at+b, according to (5.4) where y = Ln(x − c), a = −1/T and b = Ln(k).
x = ke−t/T + c (5.3)
x − c = ke−t/T => Ln(x − c) = −t/T + Ln(k) (5.4)
The problem is to find an appropriate set of k, T and c parameters through curve fitting
using some samples of (t,x) data sets. However, it can be shown that the curve fitting process
leads to a nonlinear set of equations due to appearance of x and c together in variable y [131].
Solving such a nonlinear equation needs a high processing power and is hard to implement in
practical protective relays. The other problem is that, the curve fitting process has to result in
a negative values for c. Otherwise, the approximating exponential equation of (5.3) will have
a positive DC offset. This means the approximating curve will never cross the horizontal axis,
the equation ke−t/T + c = 0 will have no real root and no real arc extinction time prediction can
be performed using this method.
In the performed research work, equation (5.4) is solved by assuming a negative value for
parameter c instead of obtaining it through solving the equations. By this assumption, the non-
linear equation becomes a linear one as c is a constant and not a variable anymore. Therefore,
the arc extinction time prediction method can be appropriately implemented in a practical relay
and also, a real output for the arc extinction time prediction method is guaranteed.
Proposed algorithm for the window-based method
(a) (b)
3
V 3 VsnD
snD
2 Approximating Approximating
exponential curve 2 exponential curve
1
Data window Data window
1
0 0
−1 −1
0.4 0.5 0.6 0.7 0.4 0.45 0.5 0.55 0.6 0.65
Time (s) Time (s)
(c) (d)
2 3
V V
snD snD
1.5 Approximating Approximating
2
exponential curve exponential curve
1
Data window 1
0.5
Data window
0
0
−0.5 −1
0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.4 0.5 0.6 0.7
Time (s) Time (s)
(e) (f)
1.5
VsnD VsnD
2
1 Approximating Approximating
exponential curve exponential curve
1
0.5 Data window
Data window
0 0
−0.5 −1
0.4 0.5 0.6 0.7 0.8 0.9 0.4 0.5 0.6 0.7 0.8 0.9
Time (s) Time (s)
Figure 5.8: V snD curves shown in (a) Figure 5.4-(e), (b) Figure 5.4-(f), (c) Figure 5.5-(e), (d)
Figure 5.5-(f), (e) Figure 5.6-(e) and (f) Figure 5.6-(f), and the associated approximating
exponentials.
V snD curves shown in Figures 5.4 to 5.6, and the associated approximating exponentials using
the window-based method are shown in Figure 5.8. As mentioned in Subsection 5.1.1, for the
time spots after the peak V snDPeak , V snD curve can be approximated by a decaying exponential
curve. So, in this research, a window of V snD data of length of six power system cycles, equal to
100ms, starting from the moment of the peak is used for the exponential curve approximation
as shown in Figure 5.8.
For the purpose of approximation, parameter c is assumed equal to −50% of V snDPeak . This
way, the term Ln(x − c) in (5.4) becomes only a function of x as c becomes just a constant.
Therefore, parameters k and T can easily be obtained by a simple linear curve fitting according
to (5.4) and (5.5). Having the parameter set of k, T and c, the root of the exponential approxi-
mating curve in (5.3) is obtained according to (5.6). This root is the zero-crossing point of the
curve which in fact, is the predicted arc extinction time. If the arc extinction prediction time is
recognized as “too long”, immediate three-phase trip instead of single-phase reclosing will be
performed despite the temporary fault type recognition.
k = eb and T = −1/a (5.5)

ke−t/T + c = 0 ⇒ t = T.Ln(k/|c|) (5.6)
5.3 Simulation Results
In this research, 225 simulated temporary fault cases have been used for performance evalua-
tion of the proposed method. The simulated cases can be categorized into three groups based
on the transposition condition of the transmission line including ideally-transposed (45 cases),
untransposed (90 cases) and partially-transposed (90 cases). Various cases are different in
terms of fault location, arc type (Arcs 1 to 3 for temporary fault cases as per Table 3.1), and
presence/absence of shunt/neutral reactor. As mentioned in the opening part of this Chapter,
the proper decision for the case of close-in fault is immediate three-phase trip signal activation
as the faulted phase voltage drops almost to zero at the relay location. So, temporary fault
inception cases at 0% distance from Side A is not considered in the studies of this Chapter.
5.3.1 Point-based method
Average arc extinction prediction time error, the difference between the arc extinction pre-
diction time by the proposed method and the real arc extinction time obtained from PSCAD
simulations, for ideally-transposed, untransposed and partially-transposed line simulated cases
are presented in Tables 5.1-5.3. The prediction time errors are available in both milliseconds
and percentage of the real arc extinction duration, the time difference between the single-phase
line isolation and the real arc extinction moments. There are two sets of data presented for
V snDT h = 0.1% and V snDT h = 1% to have access to a wider range of results. In averaging of
the results, absolute values of the time differences are used to consider both early and late pre-
dictions. Otherwise, early and late predictions may cancel out each other and in that case, the
prediction error will drop to very small values.
The key fact that must be considered regarding the results is that the predicted arc extinction
time is always longer for V snDT h = 0.1% than for V snDT h = 1% which is because of the nature of
the decaying exponential curve. Based on this fact, there are three sets of scenarios possible.
In scenario A, the time difference obtained for V snDT h = 1% is positive, i.e., the predicted arc
extinction time for V snDT h = 1% is later than the reality. In such a situation, the time difference
obtained for V snDT h = 0.1% is definitely positive and longer based on the mentioned fact. Fault
occurrence at 30% length of the untransposed line in absence of shunt reactor is an example of
such a situation (see Table 5.1).
In scenario B, the time difference for V snDT h = 0.1% is negative, means the predicted time is
earlier than the reality. What happens in this scenario is that the time difference for V snDT h = 1%
will be more negative. In this case, the absolute value of the time difference obtained for
V snDT h = 1% will be larger than for V snDT h = 0.1%. Fault occurrence at 30% length of the
untransposed line in presence of shunt reactor is an example of such a situation (see Table 5.1).
Table 5.1: Average arc extinction prediction time error of the point-based method for an
ideally-transposed/untransposed transmission line (ms).
Fault location 30% 70% 100%
Absence of V snDT h = 0.1% 132 / 307 132 / 87 67 / 115

shunt reactor V snDT h = 1% 57 / 127 57 / 59 45 / 63
Presence of V snDT h = 0.1% 46 / 83 46 / 111 117 / 194

shunt reactor V snDT h = 1% 68 / 134 68 / 124 83 / 121
Table 5.2: Average arc extinction prediction time error of the point-based method for
an ideally-transposed/untransposed transmission line in percent of the real arc extinction
time (%).
Absence of V snDT h = 0.1% 49% / 87% 49% / 28% 28% / 41%

shunt reactor V snDT h = 1% 20% / 36% 20% / 19% 10% / 23%
Presence of V snDT h = 0.1% 19% / 18% 19% / 28% 62% / 55%

shunt reactor V snDT h = 1% 29% / 29% 29% / 31% 41% / 34%
There is also a scenario C possible in which the time difference for V snDT h = 0.1% is positive
while it is negative for V snDT h = 1%. In such a scenario, the predicted time for V snDT h = 0.1%
is later than the reality while it is earlier for V snDT h = 1%. In this scenario, the proposed
method has the most accurate performance as the real arc extinction time is located between
the predicted times for V snDT h = 0.1% and V snDT h = 1%. In that case, the absolute value of
the time difference is rather small and is very close for V snDT h = 0.1% and V snDT h = 1% as
the prediction is slightly late for the first while it is slightly early for the second one. Fault
occurrence at 65% length of the partially-transposed line in presence of shunt reactor is an
example of such a situation in which the absolute values of the time differences are exactly
equal for V snDT h = 0.1% and V snDT h = 1% (see Table 5.3).
Table 5.3: Average arc extinction prediction time error of the point-based method for a
partially-transposed transmission line (ms).
Fault location 25% 50% 65% 80% 90% 100%
Absence of V snDT h = 0.1% 250 117 71 65 68 62

shunt reactor V snDT h = 1% 97 51 47 52 58 42
Presence of V snDT h = 0.1% 94 42 54 73 22 31

shunt reactor V snDT h = 1% 66 66 54 47 28 4
Table 5.4: Average arc extinction prediction time error of the point-based method for a
partially-transposed transmission line in percent of the real arc extinction time (%).
Fault location 25% 50% 65% 80% 90% 100%
Absence of V snDT h = 0.1% 83% 39% 24% 22% 23% 24%

shunt reactor V snDT h = 1% 32% 17% 16% 17% 19% 16%
Presence of V snDT h = 0.1% 31% 14% 18% 24% 7% 12%

shunt reactor V snDT h = 1% 22% 22% 18% 16% 9% 2%
Average arc extinction prediction time error for ideally-transposed, untransposed and partially-
transposed line simulated cases are shown in Figures 5.9 to 5.16. As observed, for the ideally-
transposed cases, the error values associated with the threshold value of V snDT h = 0.1% is
always higher than error values associated with V snDT h = 1%. Also, the error values related to
the two thresholds become closer for uncompensated line as the fault location becomes further
from the relay location (at 0% distance). But the gap between the error values associated with
the threshold values becomes deeper for further fault locations when the line is compensated.
For the untransposed line cases, there is no pattern observable, but for the partially-transposed
line cases there is a very clear pattern. In such cases, both the error values and also the gap
between the error values associated with the thresholds decrease for further fault locations.
Using all simulated results, average arc extinction prediction time error for ideally-transposed,
untransposed and partially-transposed line configurations are 32%, 25.8% and 31%, respec-
tively. Average prediction error of the proposed method in absence and presence of shunt
reactor are also 35.9% and 25.2% , respectively. Therefore, the proposed method has a better
performance for partially-transposed lines and in presence of shunt reactor. Finally, the pro-
posed method is able to predict the arc extinction time with error as large as 27.7% averaging
all the simulated cases.
A (V snDT h = 0.1%)
A (V snDT h = 1%)
P (V snDT h = 0.1%)
P (V snDT h = 1%)
100
50
0
30 70 100
Fault location (%)
Figure 5.9: Average arc extinction prediction time error of the point-based method for the
ideally-transposed transmission lines in Absence (A) and Presence (P) of shunt reactor
(ms).
A (V snDT h = 0.1%)
A (V snDT h = 1%)
P (V snDT h = 0.1%)
P (V snDT h = 1%)
Prediction time error (%)
60
40
20
0
30 70 100
Fault location (%)
Figure 5.10: Average arc extinction prediction time error of the point-based method for
the ideally-transposed transmission lines in Absence (A) and Presence (P) of shunt reactor
(%).
A (V snDT h = 0.1%)
A (V snDT h = 1%)
P (V snDT h = 0.1%)
P (V snDT h = 1%)
300
200
100
0
30 70 100
Fault location (%)
the untransposed transmission lines in Absence (A) and Presence (P) of shunt reactor
(ms).
A (V snDT h = 0.1%)
A (V snDT h = 1%)
P (V snDT h = 0.1%)
P (V snDT h = 1%)
80
60
40
20
0
30 70 100
Fault location (%)
the untransposed transmission lines in Absence (A) and Presence (P) of shunt reactor
(%).
Prediction time error (ms) V snDT h = 0.1% V snDT h = 1%
200
100
0
25 50 65 80 90 100
Fault location (%)
a partially-transposed transmission line in absence of shunt reactor (ms).
V snDT h = 0.1% V snDT h = 1%

80
60
40
20
0
25 50 65 80 90 100
Fault location (%)
a partially-transposed transmission line in absence of shunt reactor, in percent.

Prediction time error (ms) 100
80
60
40
20
0
25 50 65 80 90 100
Fault location (%)
a partially-transposed transmission line in presence of shunt reactor (ms).

30
20
10
0
25 50 65 80 90 100
Fault location (%)
a partially-transposed transmission line in presence of shunt reactor, in percent.
(a)
1
Vag(pu)
−1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(b)
1
|Vsn|pu
0.5
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(c)
10
0
VsnD
−10
−20
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(d)
5
Aprx. Exp.
Min
VsnD vs
VsnD
0
Peak
VsnD
−5
0.25 0.3 0.35 0.4 0.45
Time (s)
Figure 5.17: (a) Normalized instantaneous waveform, (b) magnitude and (c) the first
derivation of the faulted phase voltage, (d) the approximating exponential curve of the
765 kV faulted phase using the point-based method.
Similar to Chapter 4, performance of the proposed arc extinction time prediction method is
also evaluated using a field data, a temporary single-phase-to-ground fault in Rockport station
in American Electric Power grid on a 765 kV untransposed line with shunt reactor only at one
end [62]. Normalized instantaneous waveform, magnitude, the first derivation of the faulted
phase voltage, and the approximating exponential curve are shown in Figure 5.17. According
to the event recorder information, the fault inception time and the line single-phase isolation
time are 100ms and 177ms, respectively and waveforms say that the arc extinguishes at around
t=300ms. Therefore the arc extinction duration is 300ms-177ms=123ms. Based on the event
recorder information, reclosing occurs at t=630ms.
As shown in Figure 5.17 (d), V snD waveform is approximated by the dotted exponential
curve employing two points of V snD waveform, i.e., V snDPeak and V snDMin . Using the approx-
imating curve, the arc extinction prediction times are obtained as 404 ms and 367 ms for
V snDT h = 0.1% and V snDT h = 1%, respectively, both are much smaller than the practical re-
closing time of 630ms. Therefore, the predicted arc extinction duration is obtained as 227ms
and 190ms, respectively. Considering the real arc extinction duration of 123ms, the arc extinc-
tion time prediction error for the studied case is 84.6% and 54.5%, respectively. These error
values although are higher than the average prediction error value of 31% related to simulation
results of the untransposed line, they are still applicable for having an estimation for the arc
extinction time well in advance.
5.3.2 Window-based method
Average prediction error of the window-based prediction method, i.e., the difference between
the predicted values and PSCAD results for arc extinction duration are presented in Tables 5.5-
5.7. Arc extinction duration is the time duration between the line single-phase isolation and the
arc extinction moments. The error values, in both milliseconds and percent, are associated with
the ideally-transposed, untransposed and partially-transposed line simulated cases. Format of
the absolute error values is converted to percent form based on real arc extinction time durations
resulted from PSCAD. It should be noted that the absolute error values are used in the averaging
process so that early and late prediction results do not cancel out each other.
Regarding the overall performance of the window-based arc extinction time prediction
method, average prediction error values for ideally-transposed, untransposed and partially-
transposed line configurations are 17.7%, 18.6% and 9.9%, respectively. The average error
values for compensated and uncompensated lines are also 15.3% and 13.7%. Therefore, the
proposed method has a better performance for partially-transposed lines and in presence of
shunt reactor. Considering all the simulated cases, the proposed method is able to predict the
arc extinction time with average error as large as 14.9%.
Table 5.5: Average arc extinction prediction time error of the window-based method for
an ideally-transposed transmission line in (ms) and percent of the real arc extinction time.
Absence of (ms) 52 52 67
shunt reactor (%) 17.3 17.3 25.3
Presence of (ms) 41 41 62
shunt reactor (%) 14.1 14.1 23.6
Table 5.6: Average arc extinction prediction time error of the window-based method for
an untransposed transmission line in (ms) and percent of the real arc extinction time.
Absence of (ms) 26 40 57
shunt reactor (%) 8.2 12.7 18.7
Presence of (ms) 66 84 59
shunt reactor (%) 18.4 23.9 17.8
Table 5.7: Average arc extinction prediction time error of the window-based method for a
partially-transposed transmission line in (ms) and percent of the real arc extinction time.
Fault location 25% 50% 65% 80% 90% 100%
Absence of (ms) 36 3 22 43 49 37
shunt reactor (%) 12.6 1.1 7.8 15 17.2 13.7
Presence of (ms) 31 18 21 29 34 27
shunt reactor (%) 11 6.5 7.4 10.2 12.3 10.1
The prediction error values resulted from the window-based method in both (ms) and per-
cent are shown in bar charts of Figures 5.18 to 5.21. According to Figures 5.18 and 5.19,
comparing bars IA and IP for ideally-transposed transmission lines, the window-based method
is always more successful for shunt-compensated lines than for uncompensated ones, due to
smaller prediction errors. But the prediction is generally more accurate in absence of shunt re-
actor than in its presence for untransposed lines, considering bars UA and UP, specially when
the fault happens at the middle parts of the line. However, for partially-transposed lines, the
prediction precision depends more on the fault location than the compensation condition of the
line as seen in Figures 5.20 and 5.21. Eventually, the proposed method has a better performance
for partially-transposed lines than for ideally-transposed and untransposed lines.
IA IP UA UP
80
60
40
20
0
30 70 100
Fault location (%)
Figure 5.18: Average arc extinction prediction time error of the window-based method
for an Ideally-transposed (I) and an Untransposed (U) transmission lines in Absence (A)
and Persence (P) of shunt reactor (ms).
IA IP UA UP
20
10
0
30 70 100
Fault location (%)
for an Ideally-transposed (I) and an Untransposed (U) transmission lines in Absence (A)
and Persence (P) of shunt reactor (%).
A P

40
20
0
6525 50 80 90 100
Fault location (%)
for a partially-transposed transmission line in Absence (A) and Presence (P) of shunt
reactor (ms).
A P
15
10
0
6525 50 80 90 100
Fault location (%)
for a partially-transposed transmission line in Absence (A) and Presence (P) of shunt
reactor (%).
Performance of the method proposed for arc extinction time prediction is also evaluated
employing a filed recorded data of Rockport station which was a temporary single-phase-to-
ground fault in American Electric Power grid on a 765 kV untransposed line with shunt reactor
only at one end [62]. Normalized instantaneous waveform, magnitude, the first derivation of
the faulted phase voltage and the approximating exponential curve associated with the field
data are presented in Figure 5.22.
(a)
1
|Vag|(pu)
−1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(b)
1
|Vsn|(pu)
0.5
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(c)
40
20
VsnD
0
−20
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
VsnDPeak (d)
VsnD
Aprx. Exp.
VsnD vs
4
Zero crossing
2 Apprx. exp. curve
0
−2
0.27 0.28 0.29 0.3 0.31 0.32 0.33 0.34 0.35
Time (s)
Figure 5.22: Normalized instantaneous waveform, magnitude, the first derivation of the
faulted phase voltage and the approximating exponential curve for the real temporary
fault case using the window-based method.
According to Figure 5.22 (d), exponential curve fitting is performed using a V snD data win-
dow of length of 100ms or six power system cycles starting from the peak point. As observed,
the approximating exponential curve crosses the horizontal axis exactly at t = 300ms which
matches the event recorder information and shows a negligible prediction error. Therefore, the
proposed method has a perfect performance for the real analyzed temporary fault case.
5.3.3 Comparison of the prediction methods
The point-based prediction method is able to predict the arc extinction time with average error
as large as 27.7%. The average error of this method for ideally-transposed, untransposed and
partially-transposed line configurations is 32%, 31% and 25.8%, respectively. This method can
also predict the arc extinction time by average error of 35.9% and 25.2% for transmission lines
with and without shunt compensation, regardless of the transposition condition of the line.
Regarding the overall performance of the window-based method, average prediction error
for ideally-transposed, untransposed and partially-transposed line configurations is respectively
17.7%, 18.6% and 9.9%. The average prediction error in presence and absence of shunt reac-
tor is also 15.3% and 13.7%, respectively. Considering all the simulated cases, the proposed
method is able to predict the arc extinction time with error as large as 14.9%. Therefore,
compared to the results of the point-based method, there have been a significant improvement
achieved in the precision of the prediction for the window-based method.

Considering the performance of the two proposed methods for the case of the temporary
fault of the Rockport station, application of the point-based method has led to predicted arc
extinction moments of t=404 ms and t=367 ms using two different criteria, or prediction error
values as large as 84.6% and 54.5%, respectively, while the prediction error almost equals to
zero for the window-based method. Therefore, the window-based method has a much better
performance for the real temporary fault compared to the point-based method.
Finally, both methods has successful operation for simulation and test results while the
prediction of the window-based method are so closer to the reality. But the point-based method
is much easier to implement and needs much less processing power.
5.4 Summary
In this Chapter, two methods for the secondary arc extinction time prediction were proposed.
The proposed methods are developed for being used in making a better decision on single-phase
reclosing in transmission lines when the fault is temporary. The developed methods use the first
derivative of the faulted phase voltage magnitude. The first method, namely the point-based
method, uses only two point of the voltage derivative waveform, the peak and the minimum
of the waveform in a predetermined time period, after single-phase isolation of the line. The
second method, called the window-based method, uses a window of the voltage derivative data
after the single-phase isolation of the line. Both methods had acceptable performances while
the results obtained from the window-based method were more precise. In fact, the average
prediction error of the window-based method was 14.9% versus 27.7% error for the point-
based method. But the point-based method was much easier to implement and needed much
less processing power.

Chapter 6
Conclusions
In this research study, the first aim was to develop an adaptive method for single-phase re-
closing in high voltage transmission lines. The developed method must be able to recognize
the fault type including permanent and temporary and also, to detect arc extinction for tem-
porary faults. The method was also supposed to be local, i.e., to use only local measurement
data of the substation. The other alternative is to be communication-based and therefore, to

use measurement data associated with other substations of the power system. This was not a
subject of this research work as communication facilities are not always available. The devel-
oped method had to be comprehensive as well, meaning it must be valid for transmission lines
with different kinds of transposition conditions including ideally-transposed, untransposed and

partially-transposed, and also for different compensation conditions including shunt compen-
sated and uncompensated.
The second aim of the project was to propose a method to predict the arc extinction time
well in advance. This way, when the arc extinction process is too slow, the protection system
will have the option of issuing the three-phase trip signal instead of single-phase reclosing de-
spite the temporary fault detection. The importance of this feature is the fact that during the sys-
tem operation with one phase opened and two phases closed, negative sequence voltage/current
155
156 Chapter 6. Conclusions
components are injected to the system that can have harmful effects, such as overheating of the
generators.
In the thesis, two adaptive reclosing methods were proposed for the first part of the project.
The first proposed method, called the angle-based method, performed the reclosing task using
phase angle of the faulted phase voltage. This variable was employed for both purposes of
fault type recognition and arc extinction detection. The second proposed method, called the
phasor-based method, used the first derivative of the faulted phase voltage magnitude for the
fault type recognition, and the first derivative of both magnitude and phase angle of the faulted
phase voltage for the arc extinction detection.
It was shown in Chapter 4 that both methods had successful operations for all 550 sim-
ulated cases, including lines with ideally-transposed, untransposed and partially-transposed
configurations and also, for different transposition conditions including compensated and un-
compensated. As a comparison of the performances of the proposed methods for the simulated
cases, the phasor-based method had a better performance for permanent fault cases while per-
formance of the angle-based method was more successful for temporary fault cases at both
fault type recognition and arc extinction detection. Also, the angle-based method had a simpler
algorithm as it uses only one variable, i.e., the phase angle of the faulted phase. In addition, the
angle-based method was less sensitive to noise and fast transients as it uses the data itself and
not the derivative of the data. Therefore, for selection of the more beneficial method a trade off
must be made.
For the second part of research, two methods are developed for the arc extinction time pre-
diction purpose. Both methods are using the first derivative of the faulted phase voltage mag-
nitude. For this purpose, the derivative waveform was approximated by a decaying exponential
curve for each method and the roots of the curves would be the predicted arc extinction time.
The first method, called point-based method, uses only two points of the curve for the predic-
tion task. The second method, i.e., window-based method, uses a long window of information
157
for the prediction purpose. Both methods have local operations and therefore, no communica-
tion facilities are needed in the structures of the methods. Also, they both are comprehensive,
i.e., are valid for transmission lines with different transposition and compensation conditions.
Performances of the both methods were verified using 225 simulated temporary fault cases and
a field recorded case where both methods had successful performances. As a comparison of the
performances of the prediction methods, the output of the window-based method was closer to
reality while the point-based method is easier to implement and requires less processing power.
Suggested future works:
Regarding continuation of the research, as a very immediate suggestion, the two proposed
reclosing methods can be combined together to obtain one optimized reclosing method. The
combined method will have the best performance for both permanent and temporary fault cases.
The second suggestion is to implement the proposed methods in FPGA and other industrial
hardwares and to evaluate their performances in terms of simplicity, processing time, etc.
Nowadays, application of single-phase auto-reclosing technique is more common in dis-

tribution systems than in transmission systems. The main reason for this is the fact that in
transmission lines, the short circuit current level is much higher than the load current while
they are almost at the same level in distribution networks. Therefore, there is no consider-
able demerit in applying the reclosing technique in distribution systems. Also, it will be more
beneficial if the fault type is known before reclosing in distribution systems. Therefore, as
the concepts of the developed reclosing methods in this research study are independent from
the voltage level, one good suggestion for continuation of the research work is to apply the
developed methods to distribution systems considering the required modifications.
The reason that the proposed reclosing methods were local was the fact that in many sub-
stations, communication facilities are not available. Therefore, the forth suggestion is to add
communication facilities option to improve the performances of the methods. In this case, the
158 Chapter 6. Conclusions
improved methods will be applicable wherever the facilities exist. Although there are currently
communication-based methods presented in the literature, but the reclosing methods proposed
in this research operate almost at the same speeds. Therefore, adding communication facilities
can improve the performances even further. However, there are restrictions mentioned in differ-
ent documents such as IEC 61850 Standard regarding Communication Networks and Systems
in Substations that must be considered for the improved methods.
The proposed arc extinction time prediction methods, specially the window-based method,
are working based on some simplifications applied to the associated equations. Therefore, it is
proposed to try different techniques for solving the equations and then compare the results.
Bibliography
[1] J. D. Glover, M. S. Sarma and T. Overbye, Power Systems Analysis and Design,
Thomson-Engineering, Edition 4, 2007.
[2] IEEE Power System Relaying Committee Working Group, “Single-phase tripping and
auto re-closing of transmission lines,” IEEE Trans. Power Delivery, vol. 7, no. 1, pp.
182-192, Jan. 1992.
[3] J. Esztergalyos, J. Andrichak, D. H. Colwel1, D. C. Dawson, J. A. Jodice, T. J. Murray.
K. K. Mustaphi, G. R. Nai1, A. Politis, J. W. Pope, G. D. Rockefeller, G. P. Stranne,

D. Tziouvaras and E. O. Schweitzer, “Single phase tripping and auto reclosing of trans-
mission lines; IEEE Committee report,” IEEE Trans. Power Delivery, vol. 7, no. 1, pp.
182-192, Jan. 1992.
[4] W. Robinson and S. J. Spurgeon, “Automatic reclosing of high-voltage circuits,” Regional

Meeting of the Southwest District of the A.I.E.E., Dallas, Texas, pp. 981-987, May 7-9,
1929.
[5] IEEE Power and Energy Society, “IEEE guide for automatic reclosing of circuit breakers
for AC distribution and transmission lines,” IEEE Std C37.104-2012, 6 July 2012.
159
160 BIBLIOGRAPHY
[6] S. P. Ahn, C. H. KIM, R. K. Aggarwal and A. T. Johns, “An alternative approach to

adaptive single pole auto-reclosing in high voltage transmission systems based on variable
dead time control,” IEEE Trans. Power Delivery, vol. 16, no. 4, pp. 676-686, Oct. 2001.
[7] S. Jamali and A. Parham, “New approach to adaptive single pole autoreclosing of power
transmission lines,” IET Generation, Transmission and Distribution, vol. 4 , no. 1, pp.
115-122, Jan. 2010.
[8] IEEE Power Systems Relaying Committee “Automatic reclosing of transmission lines,”
IEEE Trans. Power Apparatous and Systems, vol. PAS-103, no. 2, pp. 234-245, Feb. 1984.
[9] M. H. Ali, T. Murata and J. Tamura, “Transient stability enhancement by fuzzy logic-
controlled SMES considering coordination with optimal reclosing of circuit breakers,”
IEEE Trans. Power Delivery, vol. 23, no. 2, pp. 631-640, May 2008.
[10] C. M. Portela, N. H. C. Santiago, 0. B. Oliveira and C. J. Dupont, “Modeling of arc

extinction in air insulation,” IEEE Trans. Electrical Insulation, vol. 27, no. 3, pp. 457-
463, June 1992.
[11] M. M. Eissa and O. P. Malik, “Experimental results of a supplementary technique for

auto-reclosing EHV/UHV transmission lines,” IEEE Trans. Power Delivery, vol. 17, no.
3, pp. 702-707, July 2002.
[12] S. Jamali and A. Parham, “New approach to adaptive single pole auto-reclosing of power
transmission lines,” IEE Proc. Generation, Transmission and Distribution, vol. 4, iss. 1,
pp. 115-122, July 2009.
[13] S. P. Le Blond and R. K. Aggarwal, “Design of adaptive autoreclosure schemes for 132
kV network with high penetration of wind-part I: real-time modeling,” IEEE Trans. Power
Delivery, vol. 27, no. 3, pp. 1055-1062, July 2012.
BIBLIOGRAPHY 161
[14] S. P. Le Blond and R. K. Aggarwal, “Design of adaptive autoreclosure schemes for 132
kV network with high penetration of wind-part II: real-time development and testing,”
IEEE Trans. Power Delivery, vol. 27, no. 3, pp. 1063-1070, July 2012.
[15] S. P. Websper, A. T. Johns, R. K. Aggarwal and R. W. Dunn, “An investigation into

breaker reclosure strategy for adaptive single pole autoreclosing,” IEE Proc. Generation,
Transmission and Distribution, vol. 142, no. 6, pp. 601-607, Nov. 1995.
[16] S. Aboreshaid, S. Aboreshaid and S. O. Faried, “Effect of adaptive single-pole reclos-
ing on the stochastic behavior of turbine-generator shaft torsional torques,” IEEE Trans.
Energy Conversion, vol. 12, no. 2, pp. 133-139, June 1998.
[17] S. O. Faried, R. Billinton,, S. Aboreshaid, and M. Fotuhi-Firuzabad, “Effect of adaptive

short time compensation on the stochastic behavior of turbine-generator shaft torsional
torques,” IEEE Trans. Energy Conversion, vol. 15, no. 3, pp. 305-311, Sept. 2000.
[18] A. M. El-Serafi and S. O. Faried, “Effect of adaptive reclosing on turbine-generator shaft

torsional torques,” IEEE Trans. Power Systems, vol. 9, no. 4, pp. 1730-1736, Nov. 1994.
[19] C. E. J. Bowler, P. G. Brown and D. N. Walker, “Evaluation of the effect of power circuit
breaker reclosing practices on turbine-generator shafts,” IEEE Trans. Power Apparatous
and Systems, vol. PAS-99, no. 5, pp. 1764-1779, Sept./Oct. 1980.
[20] E. E. Colapert and W. E. Reid, “Effects of faults and shunt reactor parameters on parallel
resonance,” IEEE Trans. Power Apparatus and Systems, vol. PAS-100, no.2, pp. 572-584,
Feb. 1981.
[21] B. R. Shperling and A. Fakheri, “Single-phase switching phase parameters for untrans-
posed EHV transmission line,” IEEE Trans. Power Apparatus and Systems, vol. PAS-98,
no.2, pp. 643-654, Mar. 1979.
162 BIBLIOGRAPHY
[22] A. Fakheri, B. J. Ware and B. R. Shperling, “Compensation scheme for single-pole

switching on untransposed transmission lines,” IEEE Trans. Power Apparatus and Sys-
tems, vol. PAS-97 , no. 4, pp.1421-1429, July 1978.
[23] M. H. Ali, T. Murata and J. Tamura, “Effect of coordination of optimal reclosing and
fuzzy controlled braking resistor on transient stability during unsuccessful reclosing,”
IEEE Trans. Power Delivery, vol. 21, no. 3, pp. 1321-1330, Aug. 2006.
[24] M. M. Eissa and O. P. Malik, “A novel approach for auto-reclosing EHV/UHV transmis-
sion lines,” IEEE Trans. Power Delivery, vol. 15, no. 3, pp. 908-912, July 2000.
[25] T. Tsuboi, J. Takami, S. Okabe, K. Aoki and Y. Yamagata, “Study on a field data of
secondary arc extinction time for large-sized transmission lines,” IEEE Trans. Dielectrics
and Electrical Insulation, vol. 20, no. 6, pp. 2277-2286, Dec. 2013.
[26] P. Mestas, M. C. Tavares and A. M. Gole, “Implementation and performance evaluation

of a reclosing method for shunt reactor-compensated transmission lines,” IEEE Trans.
Power Delivery, vol. 26, no. 2, pp. 954-962, Apr. 2011.
[27] T. Tran-Quoc, N. Hadj-Said, J. C. Sabonnadiere and R. Feuillet, “Reducing dead time for
single-phase auto-reclosing on a series-capacitor compensated transmission line,” IEEE
Trans. Power Delivery, vol. 15, no. 1, pp. 51-56, Jan. 2000.
[28] T. Tran-Quoc, H. Vo, V. Huy, H. Duong-La, L. N. Homg, and L. TrmDinh, “Single-phase
auto-reclosing of the Vietnam 500 kV system,” POWERCON Int. Conf. Power System
Technology, Beijing, China, Aug. 1998.
[29] A. J. Gonzalez, G. C. Kung, C. Raczkowski, C. W. Taylor and D. Thonn, “Effects of

single- and three-pole switching and high-speed reclosing on turbine-generator shafts and
BIBLIOGRAPHY 163
blades,” IEEE Power Apparatous and Systems, vol. PAS-103, no. 11, pp. 3218-3229, Nov.
1984.
[30] M. H. Ali, T. Murata and J. Tamura, “Transient stability enhancement by Fuzzy Logic-
controlled SMES considering coordination with optimal reclosing of circuit breakers,”
IEEE Trans. Power Systems, vol. 23, no. 2, pp. 631-640, May 2008.
[31] M. Nagpal, S. H. Manuel, B. Bell, R. Barone, C. Henville and D. Ghangass, “Field ver-
ification of secondary arc extinction logic,” IEEE Trans. Power Delivery, accpted for
publication.
[32] M. Khoroshev and V. Faybisovich, Multifactor adaptive auto-reclosing of high voltage

transmission lines US patent no. 20050195538 A1, Sept. 2005.
[33] M. C. D. Tavares and P. M. Valero, Fast three-phase reclosing method in shunt reactor
compensated transmission lines, European Patent EP 2 346 132 B1, Application number:
09820129.6, 09 Oct 2009.
[34] S. Iinuma, M. Hori and S. Shinsyou, Reclosing system for power transmission line, Patent
no. 8462474 B2, June 11, 2013.
[35] H. A. Peterson, Single phase reclosing of circuit breaker in multiphase electrical trans-
mission systems US patent no. US3505567 A, 7 Apr. 1970.
[36] S. P. Websper, A. T. Johns, R. K. Aggarwal, and R. W. Dunn, “An investigation into

breaker reclosure strategy for adaptive single pole autoreclosing,” IEE Proc. Generartion,
Transmission and Distribution, vol. 142, no. 6, pp. 601-607, Nov. 1994.
[37] D. S. Fitton, R. W. Dunn, R. K. Aggarwal, A. T. Johns, and A. Bennett, “Design and

implementation of an adaptive single pole autoreclosure technique for transmission lines
164 BIBLIOGRAPHY
using artificial neural networks,” IEEE Trans. Power Delivery, vol. 11, no. 2, pp. 748-755,
Apr. 1996.
[38] R. K. Aggarwal, A. T. Johns, R. W. Dunn and D. S. Fitton, “Neural network based adap-
tive single-pole autoreclosure technique for EHV transmission system,” IEE Proc. Gener-
artion, Transmission and Distribution, vol. 141, no. 2, pp. 155-160, 1994.
[39] F. D. Zahlay and K. S. Rama Rao, “Neuro-prony and taguchis methodology-based adap-
tive autoreclosure scheme for electric transmission systems,” IEEE Trans. Power Deliv-
ery, vol. 27, no. 2, pp. 575-582, Apr. 2012.
[40] I. K. Yu and Y. H. Song, “Wavelet analysis and neural network based adaptive single-pole
autoreclosure scheme for ehv transmission systems,” Int. Journal of Electrical Power and
Energy Systems, vol. 20, no. 7, pp. 465-474, Oct. 1998.
[41] M. H. Ali, T. Murata and J. Tamura, “Effect of coordination of optimal reclosing and
Fuzzy controlled braking resistor on transient stability during unsuccessful reclosing,”
IEEE Trans. Power Systems, vol. 21, no. 3, pp. 1321-1330, Aug. 2006.
[42] A. J. Ali, A. A. Allu and R. K. Antar, Fuzzy logic technique based single phase autore-
closing protection system of a double circuit transmission line, Proc. ICECCPCE 2013
Conf., Mosul, Iraq, Dec. 2013.
[43] B. Gireesh and M. V. Ramesh, Secondary arc extinction of single pole auto-reclosing on
EHV transmission lines with shunt compensation reactors, Proc. ICETEEEM 2012 Conf.,
Chennai, India, Dec. 2012.
[44] I. D. Kim, H. S. Cho, and J. K. Park, “Variable dead time circuit breaker auto-reclosing
scheme using artificial neural networks,” International Journal of Electrical Power and
Energy Systems, vol. 21, no. 4, pp. 269-279, May 1999.
BIBLIOGRAPHY 165
[45] S. P. Ahn, C. H. Kim, R. K. Aggarwal and A. T. Johns, “An alternative approach to

adaptive single pole auto-reclosing in high voltage transmission systems based on variable
dead time control,” IEEE Trans. Power Delivery, vol. 16, no. 4, pp. 676-686, Oct. 2001.
[46] H. A. Peterson and N. V. Dravid, “A method for reducing dead time for single-phase
reclosing in EHV transmission,” IEEE Trans. Power Apparatous and Systems, vol. PAS-
88, no. 4, pp. 286-292, Apr. 1969.
[47] R. Billinton and S. Aboreshaid, “Stochastic modelling of high-speed reclosing in proba-

bilistic transient stability studies,” IEE Proc. Generartion, Transmission and Distribution,
vol. 142, no. 4, pp. 350-354, July 1995.
[48] Y. Goda, S. Matsuda, T. Inaba and Y. Ozaki, “Insulation recovery time after fault arc
interruption for rapid auto-reclosing on UHV (lOOOkV class) transmission lines,” IEEE
Trans. Power Delivery, vol. 10, no. 2, pp. 1060-1065, Apr. 1995.
[49] G. Yaozhang, S. Fanghai and X. Yuan, “Prediction method for preventing single-phase
reclosing on permanent faults,” IEEE Trans. Power Delivery, vol. 4, no. 1, pp. 114-121,
1989.
[50] N. I. Elkalashy, H. A. Darwish, A. M. I. Taalab and M. A. Izzularab, “An adaptive single

pole auto-reclosure based on zero sequence power,” Electric Power System Research, vol.
77, no. 5-6, pp. 438-446, 2007.
[51] L. Bin, Z. Shuo, P. Crossley and B. Zhiqian, “The scheme of single-phase adaptive reclos-
ing on EHV/UHV transmission lines,” IET 9th Int. Conf. Developments in Power System
Protection (DPSP), Glasgow, Mar. 2008.
166 BIBLIOGRAPHY
[52] Z. Q. Bo, R. K. Aggarwal and A. T. Johns, “A novel technique to distinguish between

transient and permanent fault based on detection of current transients,” Proc. fourth Int.
Conf., Advances in Power System Control and Management, Hong Kong, 1997.
[53] A. Montanari, M. C. Tavares and C. M. Portela, “Adaptive single-phase auto-reclosing

based on faulted phase voltage harmonic signature,” Int. Conf. Power Systems Transients,
Kyoto, Japan, June 2009.
[54] Z. M. Radojevuc and J. R. Shin, “New digital algorithm for adaptive reclosing based on
the calculation of the faulted phase voltage total harmonic distortion factor,” IEEE Trans.
Power Delivery, vol. 22, no. 1, pp. 37-41, 2007.
[55] M. Khodadadi, M. R. Noori and S. M. Shahrtash, “A noncommunication adaptive single-

pole autoreclosure scheme based on the ACUSUM algorithm,” IEEE Trans. Power De-
livery, vol. 28, no. 4, pp. 37-41, Oct. 2013.
[56] M. R. Dadash Zadeh and R. Rubeena, “Communication-aided high-speed adaptive single-

phase reclosing,” IEEE Trans. Power Delivery, vol. 28, no. 1, pp. 499-506, Jan. 2013.
[57] Z. Q. Bo, R. K. Aggarwal, A. T. Johns, B. H. Zhang, and Y. Z. Ge., “New concept in

transmission line reclosure using high-frequency fault transients,” IEE Proc. Generartion,
Transmission and Distribution, vol. 144, no. 4, July 1997.
[58] M. E. Hamedani Golshan and N. Golbon., “Detecting secondary arc extinction time by
analyzing low frequency components of faulted phase voltage or sound phase current
waveforms,” Electrical Engineering, vol. 88, no. 2, pp. 141-148, Jan. 2006.
[59] I. K. Yu a and Y. H. Song, Development of novel adaptive singlepole autoreclosure

schemes for extra high voltage transmission systems using wavelet transform analysis,
Electric Power System Research, vol. 47, no. 1, pp. 11-19, Oct. 1998.
BIBLIOGRAPHY 167
[60] X. Lin, H. Weng, H. Liu, W. Lu, P. Liu, and Z. Bo, “A novel adaptive single-phase re-
closure scheme using dual-window transient energy ratio and mathematical morphology,”
IEEE Trans. Power Delivery, vol. 21, no. 4, pp. 1871-1877, Oct. 2006.
[61] X. Lin, H. Liu, H. Weng, P. Liu, B. Wang, and Z. Q. Bo, “A dualwindow transient energy
ratio-based adaptive single-phase reclosure criterion for EHV transmission line,” IEEE
Trans. Power Delivery, vol. 22, no. 4, pp. 2080-2086, Oct. 2007.
[62] M. R. D. Zadeh, I. Voloh, M. Kanabar and Y. Xue, “An adaptive HV transmission lines
reclosing based on voltage pattern in the complex plane,” Protective Relay Engineers,
2012 65th Ann. Conf., College Station, TX, Apr. 2012.
[63] K. M. C. Dantas, W. L. A. Neves and D. Fernandes, “An approach for controlled reclosing
of shunt-compensated transmission lines,” IEEE Trans. Power Delivery, vol. 29, no. 3, pp.
1203-1211, June 2014.
[64] D. A. Panasetsky and A. B. Osak, On the problem of shunt reactor tripping during single-
and three-phase auto-reclosing, Proc. PowerTech 2015 Conf., Eindhoven, June/July 2015.
[65] K. P. Basu1 and M. George, “Maintaining balanced three-phase load voltage during
single-phase auto-reclosing in medium voltage distribution lines,” IET Generation, Trans-

mission and Distribution, vol. 8 , Iss. 5, pp. 798-8022, May 2014.
[66] P. Choudhury,, BC Hydro Grid Operations, Operating order 7T-33, South Interior Sub-
system, Supersedes OO 7T, iss. 33, 22 November 2013.
[67] P. Choudhury, Alberta interconnected electric system protection standard, AESO Engi-
neering Standard, Revision 0, 01 Dec. 2004.
168 BIBLIOGRAPHY
[68] B. Khodabakhchian, EHV Single-pole switching: It is not only a matter of secondary arc
extinction, International Power System Transient Conf., IPST2013, Vancouver, Canada
July 18-20, 2013.
[69] J. G. Kappenman, G. A. Sweezy, V. Koschik and K. K. Mustaphi, Staged fault tests with
single phase reclosing on the Winnipeg-Twin cities 500kV interconnection, IEEE Trans.
Power Apparatus and Systems, vol. PAS-101, iss. 3, pp. 662-6734, Mar. 1982.
[70] D. Williston and D. Finney, Consequences of out-of-phase reclosing on feeders with dis-
tributed generators, 66th Annual Georgia Tech Protective Relaying Conference, Atlanta,
Georgia, Apr. 2012.
[71] C. W. Taylor, W. A. Mittlestadt, T. N. Lee, J. E. Hardy, H. Glavitsch, G. Stranne, and J.

D. Hurley, Single-pole switching for stability and reliability, IEEE Trans. Power Systems,
vol. 1, iss. 2, pp. 2526, May 1986.
[72] E. Godoy, A. Celaya, H. J. Altuve, N. Fischer, and A. Guzman, Tutorial on single-

pole tripping and reclosing, 39th Annual Western Protective Relay Conference, Spokane,
Washington, 2012.
[73] L. Edwards, J. W. Chadwick, H. A. Riesch, and L. E. Smith, Single-pole switching on

TVAs Paradise-Davidson 500-kV line design concepts and staged fault test results, IEEE
Trans. Power Apparatus and Systems, vol. PAS-90, iss. 6, pp. 24362444, Nov. 1971.
[74] Network Protection and Automation Guide, Chapter 14, p. 14-2, 2011 Alstom Grid, May
2011.
[75] J. R. Marti, Accurate modeling of frequency dependent transmission lines in electromag-

netic transients simulation,” IEEE Trans. Power Apparatus and Systems, vol. PAS-101,
no. 1, pp. 147-155, Jan. 1982.
BIBLIOGRAPHY 169
[76] EMTDC Users’ Guide, Manitoba HVDC Research Centre, 2010.
[77] S. Wenquan, Liu Yili, W. Jianhua and L. Yanbin, New algorithm for adaptive single-
phase reclosure on EHV transmission lines, Power and Energy Engineering Conference
(APPEEC), 2011 Asia-Pacific, Wuhan, 2011.
[78] L. Edwards, J. W. Chadwick, H. A. Riesch, and L. E. Smith, 500 kV single phase reclos-
ing evaluation using simplified arc model, International Journal of Emerging Technology
and Advanced Engineering, vol. 4, iss. 6, pp. 1-12, June 2014.
[79] Z. Q. Bo, R. K. Aggarwal, A. T. Johns, B. H.Zhang and Y. Z. Ge, “New concept in

transmission line reclosure using high-frequency fault transients,” IEE Proc. Generation,
Transmission and Distribution, vol. 144, no. 4, pp. 351-356, July 1997.
[80] M. Jannati, B. Vahidi, S. H. Hosseinian, and S. M. Ahadi, “A novel approach to adaptive
single phase auto-reclosing scheme for EHV transmission lines,” Electric Power Energy
Systems, vol. 33, no. 3, pp. 639-646, Mar. 2011.
[81] C. E. Weingartner, and H. E. Bonheimer, “A-C tripping and reclosing of line sectionaliz-
ing breakers,” AIEE Trans., vol. 68, pp. 1357-1359, 1949.
[82] S. Ahn, C. Kim, R. K. Aggarwal and A. T. Johns, “An alternative approach to adaptive
single pole auto-reclosing in high voltage transmission systems based on variable dead
time control,” IEEE Trans. Power Delivery, vol. 16, no. 4, pp. 676-686, Oct. 2001.
[83] M. R. Dadash Zadeh, M. Sanaye-Pasand and A. Kadivar, “Investigation of neutral reactor

performance in reducing secondary arc current,” IEEE Trans. Power Delivery, vol. 23,
no. 4, pp. 2472-2479, Oct. 2008.
170 BIBLIOGRAPHY
[84] E. W. Kimbark, “Suppression of ground-fault arcs on single-pole switched EHV lines by

shunt reactors,” IEEE Trans. Power Apparatus and Systems, vol. PAS-83, pp. 285-290,
Mar./Apr. 1964.
[85] M. C. Tavares, and C. M. Portela, “Transmission system parameters optimization sensitiv-

ity analysis of secondary arc current and recovery voltage,” IEEE Trans. Power Delivery,
vol. 19, no. 3, pp. 1464-1471, July 2004.
[86] H. Khorashadi-Zadeh and Z. Li, “Design of a novel phasor measurement unit-based trans-
mission line auto reclosing scheme,” IET Generation, Transmission and Distribution, vol.
5, iss. 8, pp. 806-813, 2011.
[87] IEEE-SA Standards Board, Power Systems Relay Committee, IEEE standard com-
mon format for trainsient data exchange (COMTRADE) for power systems, IEEE Std
C37.111¢-R2005 (Revision of IEEE Std C37.111-1999), pp. 1-56, May 2013.
[88] R. A. Ghunem, S. H. Jayaram, and E. A. Cherney, “Suppression of erosion due to dry-

band arcing under DC by alumina ri-hydrate filler in silicone rubber,” Annual Report
Conference on Electrical Insulation and Dielectric Phenomena, pp. 20-23, 2013.
[89] T. Funabashi, H. Otoguro, Y. Mizuma, L. Dube, M. Kizilcay and A. Ametani, “Influence

of fault arc characteristics on the accuracy of digital fault locators,” IEEE Trans. Power
Delivery, vol. 16, no. 2, pp. 195-199, Apr. 2001.
[90] T. Funabashi, H. Otoguro, L. Dube, M. Kizilcay and A. Ametani, “A study on fault arc
and its influence on digital fault locator performance,” IEE Conf. Developments in Power
System Protection, no. 479, Amsterdam, 2001.
BIBLIOGRAPHY 171
[91] L. Prikler, M. Kizilcay, G. Ban and P. Handl, “Improved secondary arc model based on
identification of arc parameters from staged fault test records,” 14th PSCC Conf., Sevilla,
June 2002.
[92] M. Kizilcay, G. Ban, L. Prikler and P. Handel, “Interaction of the secondary arc with the
transmission system during single-phase autoreclosure,” IEEE PowerTech 2003 Conf.,
Bologna, June 2003.
[93] C. Portela, N. Santiago, O. Oliveira and C. Dupont, “Modelling of arc extinction in air
insulation,” IEEE Trans. Dielectrics and Electrical Insulation, vol. 27, no. 3, pp. 457-463,
June 1992.
[94] V. V. Terzija, N. I. Elkalashy, G. Preston, V. Stanojevic and G. Strbac, “Detection of

arcing faults: Modelling, simulation, testing and algorithms aspects,” Proc. PowerTech
2007 Conf., Lausanne, July 2007.
[95] I. J. S. Lopes, S. H. Jayaram and E. A. Cherney, “A method for detecting the transition
from corona from water droplets to dry-band arcing on silicone rubber insulators,” IEEE
Trans. Dielectrics and Electrical Insulation, vol. 9, no. 6, pp. 964-972, Dec. 2002.
[96] Y. Goda, M. Iwata, K. Keda and S. Tanaka, “Arc voltage characteristics of high current
fault arcs in long gaps,” IEEE Trans. Power Delivery, vol. 15, no. 2, pp. 791-795, Apr.
2000.
[97] A. R. Van and C. Warrington, “Reactance relays negligibly affected by arc impedance,”
Elec. World, pp. 502-505, Sept. 19, 1931.
[98] V. V. Terzija and H. J. Koglin, “On the modeling of long arc in still air and arc resistance
calculation,” IEEE Trans. Power Delivery, vol. 19, no. 3, pp. 1012-1017, July 2004.
172 BIBLIOGRAPHY
[99] V. V. Terzija, G. Preston, M. Popov, and N. Terzija, “New static AirArc EMTP model
for long air in free air,” IEEE Trans. Power Delivery, vol. 26, no. 3, pp. 1344-1353, July
2011.
[100] M. B. Djuric and V. V. Terzija, “A new approach to the arcing faults detection for fast
autoreclosure in transmission systems,” IEEE Trans. Power Delivery, vol. 10, no. 4, pp.
1793-1798, Oct. 1995.
[101] V. V. Terzija, Z. M. Radojevic and M. B. Djuric, “A new approach for arcing faults
detection and fault distance calculation in spectral domain,” Proc. IEEE Transmission
and Distribution Conf., Los Angeles, CA, 1996.
[102] M. B. Djuric, Z. M. Radojevic and V. V. Terzija, “Time domain solution of fault distance
estimation and arcing faults detection on overhead lines,” IEEE Trans. Power Delivery,
vol. 14, no. 1, pp. 60-67, Jan. 1999.
[103] V. V. Terzija and H. J. Koglin, “Testing, modeling and simulation of long arc in still air,”
IEEE Power Engineering Society Winter Meeting, pp. 1140-1145, 2001.
[104] V. V. Terzija and H. J. Koglin, “Long arc in free air: laboratory testing, modeling, sim-
ulation and model-parameters estimation,” IEE Proc. Generation, Transmission and Dis-
tribution, vol. 149, no. 3, pp. 319-325, May 2002.
[105] V. V. Terzija and H. J. Koglin, “New approach to arc resistance calculation,” IEEE Power
Engineering Society Winter Meeting, pp. 781-787, 2001.
[106] V. V. Terzija, R. M.Ciric and H. Nouri, “Fault current calculation using hybrid compen-
sation method and new arc resistance formula,” UPEC 2004 Conf., Bristol, UK, Sept.
2004.
BIBLIOGRAPHY 173
[107] V. V. Terzija, R. Ciric and H. Nouri, “Improved fault analysis method based on a new arc
resistance formula,” IEEE Trans. Power Delivery, vol. 26, no. 1, pp. 120-126, Jan. 2011.
[108] V. V. Terzija and D. M. Dobrijevic,, “Short circuit studies in transmission networks using
improved fault model,” Proc. PowerTech 2007 Conf., Lausanne, July 2007.
[109] A. T. Johns, R. K. Aggarwal and Y. H. Song, “Improved techniques for modelling fault
arcs on faulted EHV transmission systems,” IEE Proc. Generation, Transmission and
Distribution, vol. 141, no. 2, pp. 148-154, Mar. 1994.
[110] J. Slepian and A. P. Storm, “Insulation material aids extinction of cable arcs,” Electrical
Engineering, vol. 50, iss. 12, pp. 955-957, 1931.
[111] H. C. Steiner and R. W. Strecker, “Voltage transients due to arc extinction,” AIEE Trans.
Communication and Electronics, vol. 79, no. 2, pp. 139-144, May 1960.
[112] Z. Wang, Y. Geng and Z. Liu, “Stepwise behavior of free recovery processes after dif-
fused vacuum arc extinction,” IEEE Trans. Dielectrics and Electrical Insulation, vol. 19,
no. 2, pp. 582-590, Apr. 2012.
[113] R. Landfried, T. Leblanc, M. Kirkpatrick, and P. Teste, “Assessment of the power bal-
ance at a copper cathode submitted to an electric arc by surface temperature measure-
ments and numerical modeling,” IEEE Trans. Plasma Science, vol. 40, no. 4, pp. 1205-
1216, Apr. 2012
[114] A. Allan and D. F. Amer, “The extinction of arcs in air-blast circuit-breakers,” Electrical
Engineers, vo. 94, no. 84, pp. 333-350, 1947.
[115] D. F. Alferov, G. S. Belkin and D. Y. Yevsin, “DC vacuum arc extinction in a transverse
axisymmetric magnetic field,” IEEE Trans. Plasma Science, vol. 37, no. 8, pp. 1433-1437,
Aug. 2009.
174 BIBLIOGRAPHY
[116] B. Kasztenny and V. Asaro, “CVT transient filter,” European Patent Application
EP1205015, General Electric, 2002.
[117] A. J. Fakheri, T. C. Shuter, J. M. Schneider and C. H. Shih, “Single phase switching

tests on the AEP 765 KV system-extinction time for large secondary arc currents,” IEEE
Trans. Power Apparatus and Systems, pp. 2775-2783, Aug. 1983.
[118] R. C. Van Sickle and W. E. Berkey, “Arc extinction phenomena in high voltage circuit
breakers-studied with a cathode ray oscillograph,” AIEE Trans., vo. 52, no. 3, pp. 850-
857, Sept. 1933.
[119] T. E. Browne, “Extinction of short A-C. arcs,” AIEE Trans., vol. 50, no. 4, pp. 1461-
1464, 1931.
[120] H. E. Baina, L. Xin and X. Jianyuan, “The analysis of secondary arc extinction char-
acteristics on UHV transmission lines,” 2008 Int. Conf. High Voltage Engineering and
Application, Chongqing, China, Nov. 2008.
[121] Z. Zalucki and J. Kutzner, “Initiation of electrical breakdown by ionic bombardment in

a vacuum gap after arc extinction,” IEEE Trans. Plasma Science, vol. PS-13, no. 5, pp.
315-320, Oct. 1985.
[122] A. J. Fakheri, T. C. Shuter, T. C. Shuter and C. H. Shih, “Single phase switching tests
on the AEP 765 kV system extinction time for large secondary arc currents,” IEEE Trans.
Power Apparatus and Systems, vol. PAS-102, no. 8, pp. 2775-2783, Aug. 1983.
[123] B. Miedzinski and T. Zolnierczyk, “Influence of dielectric liquid on arc extinction in

REED switches,” IEEE Trans. Electrical Insulation, vol. EI-1s, no. 5, pp. 404-412, Oct.
1980.
BIBLIOGRAPHY 175
[124] A. V. Batrakov, B. Juttner,D. I. Proskurovsky and E. L. Pryadko, “Light emission of

droplet spots at vacuum arc and after arc extinction,” IEEE Trans. Plasma Science, vol.
33, no. 5, pp. 1476-1480, Oct. 2005.
[125] G. A. Farrall, “Decay of residual plasma in a vacuum gap after forced extinction of a
250-ampere arc,” Proc. IEEE, vol. 56, no. 12, pp. 2137-2145, Dec. 1968.
[126] H. K. Cho, E. W. Lee and S. S. Lim, “Estimating factor for improving arc extinction ca-
pability of DC magnetic switch,” IEEE Asia Pacific Transmission and Distribution Con-
ference and Exhibition, vol. 3, pp. 1770-1773, 2002.
[127] A. V. Budin, A. A. Bogomaz, V. A. Kolikov, Ph. G. Rutberg and A. F. Sawateev. “Mul-

tipulse discharge in the chamber of electric discharge launcher,” IEEE Trans. Magnetics,
vol. 35, no. 1, pp. 189-191, Jan. 1999.
[128] Y. Goda, S. Matsuda, T. Inaba and Y. Ozaki, “Forced extinction characteristics of sec-
ondary arc on UHV (1000 kV class) transmission lines,” IEEE Trans. Power Delivery,
vol. 8, no. 3, pp. 1322-1330, July 1993.
[129] R. Peter and P. Smeets, “The origin of current chopping in vacuum arcs,” IEEE Trans.
Plasma Science, vol. 17, no. 2, pp. 303-310, Apr. 1989.
[130] E. O. Schweitzer and D. Hou, “Filtering for protective relays,” 19th Ann. Western Pro-
tection Relay Conf., Spokane, WA, USA, 1992.
[131] A. Croft, M. Hargreaves and R. Davison, Engineering Mathematics, Pearson Education,

July 2012.
176
Curriculum Vitae
Name: Farzad Zhalefar
Post-secondary Zanjan University

Education and Zanjan, Iran
Degrees: 1999-2003 B.A.
University of Tehran
Tehran, Iran
2003-2006 M.Sc.

London, Ontario, Canada
2010-2015 Ph.D.
Honours and Rank 2nd in Scientific Exam

Awards: Ministry of Education, Zanjan, Iran
1992
Rank 2nd amongst electrical engineering students

Zanjan University, Zanjan, Iran
2003
Rank 1st at entrance level amongst power engineering students

University of Tehran, Tehran, Iran
2003
Related Work Graduate Research/Teaching Assistant

Experience The University of Western Ontario
2010-2015
Power Transformer Designer

Technical Office of Power Transformers’ Factory,
Iran-Transfo Co., Zanjan, Iran
2007-2010
College Lecturer
Mahdavi Technical Colledge, Zanjan, Iran
2007-2008
Researcher
Iran Transformer Research Institute
Iran-Transfo Co., Zanjan, Iran
2005-2007
177
Publications:
• F. Zhalefar, M.R. Dadashzadeh and T.S. Sidhu, “A high-speed adaptive single-

phase reclosing technique based on local voltage phasors,” Accepted for
publication, IEEE Transactions on Power Delivery.
• F. Zhalefar, M.R. Dadashzadeh and T.S. Sidhu, Prediction of arc extinction time
for single-phase reclosing of transmission lines, Accepted at PES General Meeting
2015, July 2015, Denver, Colorado.
• F. Zhalefar and M. Sanaye-Pasand, “A new fuzzy-logic-based extended blocking

scheme for differential protection of power transformers,” Electric Power
Components and Systems, vol. 38, iss. 6, Apr. 2010, pp. 675-694.
• F. Zhalefar, M. Kalantari and J. Faiz, Studying the effect of location of tap-

changer switch on flux density of magnetic core, International Colloquium
Transformer Research and Asset Management, Nov. 2009, Croatia.
• F. Zhalefar and M. Sanaye-Pasand, Calculation of no-load loss in power

transformers with five-limbed magnetic core; a case study, International
Colloquium Transformer Research and Asset Management, Nov. 2009, Croatia.
• G.A. Zand, M. Moghadami and F. Zhalefar, Report of diagnosis and repair of 220
MVA power transformer of Mobarakeh Steel Corporation’s Substation, PSC 2008,
Nov. 2008, Tehran, Iran (in Persian/Farsi).
• M. Sanaye-Pasand and F. Zhalefar, Application of differential and REF methods

for protection of power transformers, a comparative study, International
Conference on Power Systems, Dec. 2007, Bangalore, India.
• F. Zhalefar and M. Sanaye-Pasand, Studying effect of location and resistance of

inter-turn fault on fault current in power transformers, UPEC 2007, Sept. 2007,
Brighton, UK.
• F. Zhalefar and M. Sanaye-Pasand, Transformer winding detailed modeling for

protection studies, IEEE Power India 2006, April 2006, New Delhi.
• F. Zhalefar and M. Sanaye-Pasand, Transformer core modeling as applied to slow

transients and protective studies, IEEE Power India 2006, April 2006, New Delhi.

Adaptive Single-Phase Reclosing in Transmission Lines Adaptive Single-Phase Reclosing in Transmission Lines

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Adaptive Single-Phase Reclosing in Transmission Lines Adaptive Single-Phase Reclosing in Transmission Lines

Uploaded by

Copyright:

Available Formats

View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

Electronic Thesis and Dissertation Repository

Adaptive Single-Phase Reclosing in Transmission Lines

Graduate Program in Electrical and Computer Engineering

Follow this and additional works at: https://ir.lib.uwo.ca/etd

(Thesis format: Monograph)

Graduate Program in Electrical and Computer Engineering

A thesis submitted in partial fulfillment

of the requirements for the degree of

The School of Graduate and Postdoctoral Studies

The University of Western Ontario

London, Ontario, Canada

c Farzad Zhalefar 2015

time well in advance and with acceptable precision.

considerable improvements are observed.

Keywords: Adaptive protection, arc extinction, single-phase reclosing, tranmission line

the helps and supports from my dearest family.

List of Tables xviii

1.2 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Reclosing in Transmission Lines 11

2.2 Single-Phase Reclosing in Canada . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.2 Single-shot vs. multiple-shot . . . . . . . . . . . . . . . . . . . . . . . 17

2.5 Application Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.6 IEEE Guides for Auto-Reclosing . . . . . . . . . . . . . . . . . . . . . . . . . 29

2.6.1 Fundamentals and applications . . . . . . . . . . . . . . . . . . . . . . 29

2.6.2 Auto-reclosing for transmission systems . . . . . . . . . . . . . . . . . 32

2.6.4 Auto-reclosing settings . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.6.5 Auto-reclosing blocking . . . . . . . . . . . . . . . . . . . . . . . . . 37

2.6.6 Operation considerations . . . . . . . . . . . . . . . . . . . . . . . . . 38

3 Modeling and Behavior Analysis During Single-phase Reclosing 54

3.1 Modeling of the Studied System . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.1.1 Power system modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.1.2 Arc modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.1.3 Protective relay modeling . . . . . . . . . . . . . . . . . . . . . . . . . 64

3.2 Behavior Analysis of the System During Single-Phase Reclosing . . . . . . . . 66

3.2.1 Transposed line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Absence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 66

Presence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 69

3.2.2 Untransposed line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

Absence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 72

Presence of shunt reactor . . . . . . . . . . . . . . . . . . . . . . . . . 73

4 Fault Type Recognition and Arc Extinction Detection 76

4.1 Fundamentals of the Reclosing Methods . . . . . . . . . . . . . . . . . . . . . 77

4.1.1 Angle-based analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.1.2 Phasor-based analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

4.2 Proposed Adaptive Reclosing Algorithms . . . . . . . . . . . . . . . . . . . . 100

4.2.1 Angle-based algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Fault type recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Arc extinction detection . . . . . . . . . . . . . . . . . . . . . . . . . 102

4.2.2 Phasor-based algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Fault type recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Arc extinction detection . . . . . . . . . . . . . . . . . . . . . . . . . 105

4.3 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

4.3.1 Angle-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

4.3.2 Phasor-based method . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

4.3.3 Comparison of the reclosing methods . . . . . . . . . . . . . . . . . . 124

4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

5 Arc Extinction Time Prediction 127

5.1 Fundamentals of The Proposed Prediction Methods . . . . . . . . . . . . . . . 128

5.1.1 Analysis of the faulted phase voltage waveforms . . . . . . . . . . . . 128

5.2 Proposed Prediction Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 132